Upload folder using huggingface_hub
Browse files- best_model/config.json +81 -0
- best_model/id2label.json +28 -0
- best_model/label2id.json +28 -0
- best_model/model.safetensors +3 -0
- best_model/special_tokens_map.json +7 -0
- best_model/tokenizer.json +0 -0
- best_model/tokenizer_config.json +59 -0
- best_model/training_args.bin +3 -0
- best_model/vocab.txt +0 -0
- checkpoint-11204/config.json +81 -0
- checkpoint-11204/model.safetensors +3 -0
- checkpoint-11204/optimizer.pt +3 -0
- checkpoint-11204/rng_state.pth +3 -0
- checkpoint-11204/scaler.pt +3 -0
- checkpoint-11204/scheduler.pt +3 -0
- checkpoint-11204/special_tokens_map.json +7 -0
- checkpoint-11204/tokenizer.json +0 -0
- checkpoint-11204/tokenizer_config.json +59 -0
- checkpoint-11204/trainer_state.json +1651 -0
- checkpoint-11204/training_args.bin +3 -0
- checkpoint-11204/vocab.txt +0 -0
- checkpoint-14005/config.json +81 -0
- checkpoint-14005/model.safetensors +3 -0
- checkpoint-14005/optimizer.pt +3 -0
- checkpoint-14005/rng_state.pth +3 -0
- checkpoint-14005/scaler.pt +3 -0
- checkpoint-14005/scheduler.pt +3 -0
- checkpoint-14005/special_tokens_map.json +7 -0
- checkpoint-14005/tokenizer.json +0 -0
- checkpoint-14005/tokenizer_config.json +59 -0
- checkpoint-14005/trainer_state.json +2053 -0
- checkpoint-14005/training_args.bin +3 -0
- checkpoint-14005/vocab.txt +0 -0
- metrics.json +21 -0
- split_test.csv +0 -0
- split_train.csv +0 -0
- split_valid.csv +0 -0
best_model/config.json
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "Aleppo",
|
| 14 |
+
"1": "Alexandria",
|
| 15 |
+
"2": "Algiers",
|
| 16 |
+
"3": "Amman",
|
| 17 |
+
"4": "Aswan",
|
| 18 |
+
"5": "Baghdad",
|
| 19 |
+
"6": "Basra",
|
| 20 |
+
"7": "Beirut",
|
| 21 |
+
"8": "Benghazi",
|
| 22 |
+
"9": "Cairo",
|
| 23 |
+
"10": "Damascus",
|
| 24 |
+
"11": "Doha",
|
| 25 |
+
"12": "Fes",
|
| 26 |
+
"13": "Jeddah",
|
| 27 |
+
"14": "Jerusalem",
|
| 28 |
+
"15": "Khartoum",
|
| 29 |
+
"16": "MSA",
|
| 30 |
+
"17": "Mosul",
|
| 31 |
+
"18": "Muscat",
|
| 32 |
+
"19": "Rabat",
|
| 33 |
+
"20": "Riyadh",
|
| 34 |
+
"21": "Salt",
|
| 35 |
+
"22": "Sanaa",
|
| 36 |
+
"23": "Sfax",
|
| 37 |
+
"24": "Tripoli",
|
| 38 |
+
"25": "Tunis"
|
| 39 |
+
},
|
| 40 |
+
"initializer_range": 0.02,
|
| 41 |
+
"intermediate_size": 3072,
|
| 42 |
+
"label2id": {
|
| 43 |
+
"Aleppo": 0,
|
| 44 |
+
"Alexandria": 1,
|
| 45 |
+
"Algiers": 2,
|
| 46 |
+
"Amman": 3,
|
| 47 |
+
"Aswan": 4,
|
| 48 |
+
"Baghdad": 5,
|
| 49 |
+
"Basra": 6,
|
| 50 |
+
"Beirut": 7,
|
| 51 |
+
"Benghazi": 8,
|
| 52 |
+
"Cairo": 9,
|
| 53 |
+
"Damascus": 10,
|
| 54 |
+
"Doha": 11,
|
| 55 |
+
"Fes": 12,
|
| 56 |
+
"Jeddah": 13,
|
| 57 |
+
"Jerusalem": 14,
|
| 58 |
+
"Khartoum": 15,
|
| 59 |
+
"MSA": 16,
|
| 60 |
+
"Mosul": 17,
|
| 61 |
+
"Muscat": 18,
|
| 62 |
+
"Rabat": 19,
|
| 63 |
+
"Riyadh": 20,
|
| 64 |
+
"Salt": 21,
|
| 65 |
+
"Sanaa": 22,
|
| 66 |
+
"Sfax": 23,
|
| 67 |
+
"Tripoli": 24,
|
| 68 |
+
"Tunis": 25
|
| 69 |
+
},
|
| 70 |
+
"layer_norm_eps": 1e-12,
|
| 71 |
+
"max_position_embeddings": 512,
|
| 72 |
+
"model_type": "bert",
|
| 73 |
+
"num_attention_heads": 12,
|
| 74 |
+
"num_hidden_layers": 12,
|
| 75 |
+
"pad_token_id": 0,
|
| 76 |
+
"position_embedding_type": "absolute",
|
| 77 |
+
"transformers_version": "4.56.2",
|
| 78 |
+
"type_vocab_size": 2,
|
| 79 |
+
"use_cache": true,
|
| 80 |
+
"vocab_size": 30000
|
| 81 |
+
}
|
best_model/id2label.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"0": "Aleppo",
|
| 3 |
+
"1": "Alexandria",
|
| 4 |
+
"2": "Algiers",
|
| 5 |
+
"3": "Amman",
|
| 6 |
+
"4": "Aswan",
|
| 7 |
+
"5": "Baghdad",
|
| 8 |
+
"6": "Basra",
|
| 9 |
+
"7": "Beirut",
|
| 10 |
+
"8": "Benghazi",
|
| 11 |
+
"9": "Cairo",
|
| 12 |
+
"10": "Damascus",
|
| 13 |
+
"11": "Doha",
|
| 14 |
+
"12": "Fes",
|
| 15 |
+
"13": "Jeddah",
|
| 16 |
+
"14": "Jerusalem",
|
| 17 |
+
"15": "Khartoum",
|
| 18 |
+
"16": "MSA",
|
| 19 |
+
"17": "Mosul",
|
| 20 |
+
"18": "Muscat",
|
| 21 |
+
"19": "Rabat",
|
| 22 |
+
"20": "Riyadh",
|
| 23 |
+
"21": "Salt",
|
| 24 |
+
"22": "Sanaa",
|
| 25 |
+
"23": "Sfax",
|
| 26 |
+
"24": "Tripoli",
|
| 27 |
+
"25": "Tunis"
|
| 28 |
+
}
|
best_model/label2id.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"Aleppo": 0,
|
| 3 |
+
"Alexandria": 1,
|
| 4 |
+
"Algiers": 2,
|
| 5 |
+
"Amman": 3,
|
| 6 |
+
"Aswan": 4,
|
| 7 |
+
"Baghdad": 5,
|
| 8 |
+
"Basra": 6,
|
| 9 |
+
"Beirut": 7,
|
| 10 |
+
"Benghazi": 8,
|
| 11 |
+
"Cairo": 9,
|
| 12 |
+
"Damascus": 10,
|
| 13 |
+
"Doha": 11,
|
| 14 |
+
"Fes": 12,
|
| 15 |
+
"Jeddah": 13,
|
| 16 |
+
"Jerusalem": 14,
|
| 17 |
+
"Khartoum": 15,
|
| 18 |
+
"MSA": 16,
|
| 19 |
+
"Mosul": 17,
|
| 20 |
+
"Muscat": 18,
|
| 21 |
+
"Rabat": 19,
|
| 22 |
+
"Riyadh": 20,
|
| 23 |
+
"Salt": 21,
|
| 24 |
+
"Sanaa": 22,
|
| 25 |
+
"Sfax": 23,
|
| 26 |
+
"Tripoli": 24,
|
| 27 |
+
"Tunis": 25
|
| 28 |
+
}
|
best_model/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e81f9b3ea982c64a859dd00899740789032f2821f6f53fb2d17e3406df66988a
|
| 3 |
+
size 436428880
|
best_model/special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
best_model/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
best_model/tokenizer_config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"full_tokenizer_file": null,
|
| 50 |
+
"mask_token": "[MASK]",
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"never_split": null,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"sep_token": "[SEP]",
|
| 55 |
+
"strip_accents": null,
|
| 56 |
+
"tokenize_chinese_chars": true,
|
| 57 |
+
"tokenizer_class": "BertTokenizer",
|
| 58 |
+
"unk_token": "[UNK]"
|
| 59 |
+
}
|
best_model/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:15055f48195ad08339f81e25f1ac7a0da4c4b8b3553f36d6a79688713db2f8bb
|
| 3 |
+
size 5777
|
best_model/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-11204/config.json
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "Aleppo",
|
| 14 |
+
"1": "Alexandria",
|
| 15 |
+
"2": "Algiers",
|
| 16 |
+
"3": "Amman",
|
| 17 |
+
"4": "Aswan",
|
| 18 |
+
"5": "Baghdad",
|
| 19 |
+
"6": "Basra",
|
| 20 |
+
"7": "Beirut",
|
| 21 |
+
"8": "Benghazi",
|
| 22 |
+
"9": "Cairo",
|
| 23 |
+
"10": "Damascus",
|
| 24 |
+
"11": "Doha",
|
| 25 |
+
"12": "Fes",
|
| 26 |
+
"13": "Jeddah",
|
| 27 |
+
"14": "Jerusalem",
|
| 28 |
+
"15": "Khartoum",
|
| 29 |
+
"16": "MSA",
|
| 30 |
+
"17": "Mosul",
|
| 31 |
+
"18": "Muscat",
|
| 32 |
+
"19": "Rabat",
|
| 33 |
+
"20": "Riyadh",
|
| 34 |
+
"21": "Salt",
|
| 35 |
+
"22": "Sanaa",
|
| 36 |
+
"23": "Sfax",
|
| 37 |
+
"24": "Tripoli",
|
| 38 |
+
"25": "Tunis"
|
| 39 |
+
},
|
| 40 |
+
"initializer_range": 0.02,
|
| 41 |
+
"intermediate_size": 3072,
|
| 42 |
+
"label2id": {
|
| 43 |
+
"Aleppo": 0,
|
| 44 |
+
"Alexandria": 1,
|
| 45 |
+
"Algiers": 2,
|
| 46 |
+
"Amman": 3,
|
| 47 |
+
"Aswan": 4,
|
| 48 |
+
"Baghdad": 5,
|
| 49 |
+
"Basra": 6,
|
| 50 |
+
"Beirut": 7,
|
| 51 |
+
"Benghazi": 8,
|
| 52 |
+
"Cairo": 9,
|
| 53 |
+
"Damascus": 10,
|
| 54 |
+
"Doha": 11,
|
| 55 |
+
"Fes": 12,
|
| 56 |
+
"Jeddah": 13,
|
| 57 |
+
"Jerusalem": 14,
|
| 58 |
+
"Khartoum": 15,
|
| 59 |
+
"MSA": 16,
|
| 60 |
+
"Mosul": 17,
|
| 61 |
+
"Muscat": 18,
|
| 62 |
+
"Rabat": 19,
|
| 63 |
+
"Riyadh": 20,
|
| 64 |
+
"Salt": 21,
|
| 65 |
+
"Sanaa": 22,
|
| 66 |
+
"Sfax": 23,
|
| 67 |
+
"Tripoli": 24,
|
| 68 |
+
"Tunis": 25
|
| 69 |
+
},
|
| 70 |
+
"layer_norm_eps": 1e-12,
|
| 71 |
+
"max_position_embeddings": 512,
|
| 72 |
+
"model_type": "bert",
|
| 73 |
+
"num_attention_heads": 12,
|
| 74 |
+
"num_hidden_layers": 12,
|
| 75 |
+
"pad_token_id": 0,
|
| 76 |
+
"position_embedding_type": "absolute",
|
| 77 |
+
"transformers_version": "4.56.2",
|
| 78 |
+
"type_vocab_size": 2,
|
| 79 |
+
"use_cache": true,
|
| 80 |
+
"vocab_size": 30000
|
| 81 |
+
}
|
checkpoint-11204/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eabceee2ba5d0c33f88da2de4cd15111c4517bbac407e40bb9331a0048794710
|
| 3 |
+
size 436428880
|
checkpoint-11204/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ec87b897a85eab1cdf482c0f27f7d83b56195fee8d8c39bb4ceeaea93dc334a
|
| 3 |
+
size 872982027
|
checkpoint-11204/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d41129f351a0a7e38486e07fd4388394130e6d7460b56cb0e29eae6b3a31a28
|
| 3 |
+
size 14645
|
checkpoint-11204/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:45f7d9ee07d10f01f34441b5d16c653141a554d2f89fd2222ea20ca9f731de92
|
| 3 |
+
size 1383
|
checkpoint-11204/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8806d222011c06d6f0cc9a6e08b1647a37bebec5527e67d901a717059cd88ae
|
| 3 |
+
size 1465
|
checkpoint-11204/special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
checkpoint-11204/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-11204/tokenizer_config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"full_tokenizer_file": null,
|
| 50 |
+
"mask_token": "[MASK]",
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"never_split": null,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"sep_token": "[SEP]",
|
| 55 |
+
"strip_accents": null,
|
| 56 |
+
"tokenize_chinese_chars": true,
|
| 57 |
+
"tokenizer_class": "BertTokenizer",
|
| 58 |
+
"unk_token": "[UNK]"
|
| 59 |
+
}
|
checkpoint-11204/trainer_state.json
ADDED
|
@@ -0,0 +1,1651 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 11204,
|
| 3 |
+
"best_metric": 0.5141377387367236,
|
| 4 |
+
"best_model_checkpoint": "camelbert_madar_all/checkpoint-11204",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 11204,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.01785076758300607,
|
| 14 |
+
"grad_norm": 12.478485107421875,
|
| 15 |
+
"learning_rate": 1.9930024991074616e-05,
|
| 16 |
+
"loss": 3.29,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.03570153516601214,
|
| 21 |
+
"grad_norm": 8.7147798538208,
|
| 22 |
+
"learning_rate": 1.9858621920742592e-05,
|
| 23 |
+
"loss": 3.0927,
|
| 24 |
+
"step": 100
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.053552302749018205,
|
| 28 |
+
"grad_norm": 9.888656616210938,
|
| 29 |
+
"learning_rate": 1.978721885041057e-05,
|
| 30 |
+
"loss": 2.8678,
|
| 31 |
+
"step": 150
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.07140307033202428,
|
| 35 |
+
"grad_norm": 11.028563499450684,
|
| 36 |
+
"learning_rate": 1.9715815780078543e-05,
|
| 37 |
+
"loss": 2.6757,
|
| 38 |
+
"step": 200
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.08925383791503035,
|
| 42 |
+
"grad_norm": 16.397783279418945,
|
| 43 |
+
"learning_rate": 1.9644412709746522e-05,
|
| 44 |
+
"loss": 2.593,
|
| 45 |
+
"step": 250
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.10710460549803641,
|
| 49 |
+
"grad_norm": 14.620532989501953,
|
| 50 |
+
"learning_rate": 1.9573009639414498e-05,
|
| 51 |
+
"loss": 2.5221,
|
| 52 |
+
"step": 300
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.12495537308104249,
|
| 56 |
+
"grad_norm": 18.156570434570312,
|
| 57 |
+
"learning_rate": 1.9501606569082473e-05,
|
| 58 |
+
"loss": 2.4798,
|
| 59 |
+
"step": 350
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.14280614066404856,
|
| 63 |
+
"grad_norm": 16.055646896362305,
|
| 64 |
+
"learning_rate": 1.943020349875045e-05,
|
| 65 |
+
"loss": 2.4239,
|
| 66 |
+
"step": 400
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.16065690824705461,
|
| 70 |
+
"grad_norm": 16.83954620361328,
|
| 71 |
+
"learning_rate": 1.9358800428418424e-05,
|
| 72 |
+
"loss": 2.3324,
|
| 73 |
+
"step": 450
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.1785076758300607,
|
| 77 |
+
"grad_norm": 11.909624099731445,
|
| 78 |
+
"learning_rate": 1.92873973580864e-05,
|
| 79 |
+
"loss": 2.3261,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.19635844341306677,
|
| 84 |
+
"grad_norm": 11.35456371307373,
|
| 85 |
+
"learning_rate": 1.9215994287754376e-05,
|
| 86 |
+
"loss": 2.1992,
|
| 87 |
+
"step": 550
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.21420921099607282,
|
| 91 |
+
"grad_norm": 13.982519149780273,
|
| 92 |
+
"learning_rate": 1.914459121742235e-05,
|
| 93 |
+
"loss": 2.1815,
|
| 94 |
+
"step": 600
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.2320599785790789,
|
| 98 |
+
"grad_norm": 13.58781623840332,
|
| 99 |
+
"learning_rate": 1.9073188147090327e-05,
|
| 100 |
+
"loss": 2.1739,
|
| 101 |
+
"step": 650
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.24991074616208497,
|
| 105 |
+
"grad_norm": 12.896429061889648,
|
| 106 |
+
"learning_rate": 1.9001785076758302e-05,
|
| 107 |
+
"loss": 2.1063,
|
| 108 |
+
"step": 700
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.26776151374509105,
|
| 112 |
+
"grad_norm": 12.02055549621582,
|
| 113 |
+
"learning_rate": 1.8930382006426278e-05,
|
| 114 |
+
"loss": 2.1043,
|
| 115 |
+
"step": 750
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.28561228132809713,
|
| 119 |
+
"grad_norm": 13.458044052124023,
|
| 120 |
+
"learning_rate": 1.8858978936094254e-05,
|
| 121 |
+
"loss": 2.1073,
|
| 122 |
+
"step": 800
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.30346304891110315,
|
| 126 |
+
"grad_norm": 19.288684844970703,
|
| 127 |
+
"learning_rate": 1.878757586576223e-05,
|
| 128 |
+
"loss": 2.0457,
|
| 129 |
+
"step": 850
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.32131381649410923,
|
| 133 |
+
"grad_norm": 12.411849021911621,
|
| 134 |
+
"learning_rate": 1.8716172795430205e-05,
|
| 135 |
+
"loss": 1.9664,
|
| 136 |
+
"step": 900
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.3391645840771153,
|
| 140 |
+
"grad_norm": 13.870584487915039,
|
| 141 |
+
"learning_rate": 1.864476972509818e-05,
|
| 142 |
+
"loss": 1.9998,
|
| 143 |
+
"step": 950
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.3570153516601214,
|
| 147 |
+
"grad_norm": 9.050592422485352,
|
| 148 |
+
"learning_rate": 1.8573366654766156e-05,
|
| 149 |
+
"loss": 2.0137,
|
| 150 |
+
"step": 1000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.37486611924312746,
|
| 154 |
+
"grad_norm": 10.100810050964355,
|
| 155 |
+
"learning_rate": 1.850196358443413e-05,
|
| 156 |
+
"loss": 1.9402,
|
| 157 |
+
"step": 1050
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.39271688682613354,
|
| 161 |
+
"grad_norm": 15.232287406921387,
|
| 162 |
+
"learning_rate": 1.8430560514102107e-05,
|
| 163 |
+
"loss": 1.8844,
|
| 164 |
+
"step": 1100
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.4105676544091396,
|
| 168 |
+
"grad_norm": 18.118881225585938,
|
| 169 |
+
"learning_rate": 1.8359157443770083e-05,
|
| 170 |
+
"loss": 1.9186,
|
| 171 |
+
"step": 1150
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.42841842199214564,
|
| 175 |
+
"grad_norm": 19.386072158813477,
|
| 176 |
+
"learning_rate": 1.8287754373438062e-05,
|
| 177 |
+
"loss": 1.8201,
|
| 178 |
+
"step": 1200
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.4462691895751517,
|
| 182 |
+
"grad_norm": 15.369712829589844,
|
| 183 |
+
"learning_rate": 1.8216351303106034e-05,
|
| 184 |
+
"loss": 1.8765,
|
| 185 |
+
"step": 1250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.4641199571581578,
|
| 189 |
+
"grad_norm": 16.337072372436523,
|
| 190 |
+
"learning_rate": 1.814494823277401e-05,
|
| 191 |
+
"loss": 1.7933,
|
| 192 |
+
"step": 1300
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.48197072474116387,
|
| 196 |
+
"grad_norm": 12.64062786102295,
|
| 197 |
+
"learning_rate": 1.8073545162441985e-05,
|
| 198 |
+
"loss": 1.8877,
|
| 199 |
+
"step": 1350
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.49982149232416995,
|
| 203 |
+
"grad_norm": 9.410002708435059,
|
| 204 |
+
"learning_rate": 1.800214209210996e-05,
|
| 205 |
+
"loss": 1.8393,
|
| 206 |
+
"step": 1400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.517672259907176,
|
| 210 |
+
"grad_norm": 14.48257827758789,
|
| 211 |
+
"learning_rate": 1.793073902177794e-05,
|
| 212 |
+
"loss": 1.7939,
|
| 213 |
+
"step": 1450
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.5355230274901821,
|
| 217 |
+
"grad_norm": 14.2874116897583,
|
| 218 |
+
"learning_rate": 1.7859335951445912e-05,
|
| 219 |
+
"loss": 1.8296,
|
| 220 |
+
"step": 1500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.5533737950731882,
|
| 224 |
+
"grad_norm": 13.599530220031738,
|
| 225 |
+
"learning_rate": 1.778793288111389e-05,
|
| 226 |
+
"loss": 1.8267,
|
| 227 |
+
"step": 1550
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.5712245626561943,
|
| 231 |
+
"grad_norm": 15.281228065490723,
|
| 232 |
+
"learning_rate": 1.7716529810781867e-05,
|
| 233 |
+
"loss": 1.8124,
|
| 234 |
+
"step": 1600
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.5890753302392003,
|
| 238 |
+
"grad_norm": 16.20423126220703,
|
| 239 |
+
"learning_rate": 1.7645126740449842e-05,
|
| 240 |
+
"loss": 1.7258,
|
| 241 |
+
"step": 1650
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.6069260978222063,
|
| 245 |
+
"grad_norm": 17.073945999145508,
|
| 246 |
+
"learning_rate": 1.7573723670117818e-05,
|
| 247 |
+
"loss": 1.7866,
|
| 248 |
+
"step": 1700
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.6247768654052124,
|
| 252 |
+
"grad_norm": 12.85055160522461,
|
| 253 |
+
"learning_rate": 1.750232059978579e-05,
|
| 254 |
+
"loss": 1.7171,
|
| 255 |
+
"step": 1750
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.6426276329882185,
|
| 259 |
+
"grad_norm": 26.679119110107422,
|
| 260 |
+
"learning_rate": 1.743091752945377e-05,
|
| 261 |
+
"loss": 1.7145,
|
| 262 |
+
"step": 1800
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.6604784005712245,
|
| 266 |
+
"grad_norm": 20.80165672302246,
|
| 267 |
+
"learning_rate": 1.7359514459121744e-05,
|
| 268 |
+
"loss": 1.7033,
|
| 269 |
+
"step": 1850
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.6783291681542306,
|
| 273 |
+
"grad_norm": 12.617931365966797,
|
| 274 |
+
"learning_rate": 1.728811138878972e-05,
|
| 275 |
+
"loss": 1.6911,
|
| 276 |
+
"step": 1900
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.6961799357372367,
|
| 280 |
+
"grad_norm": 12.220250129699707,
|
| 281 |
+
"learning_rate": 1.7216708318457696e-05,
|
| 282 |
+
"loss": 1.7364,
|
| 283 |
+
"step": 1950
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.7140307033202428,
|
| 287 |
+
"grad_norm": 12.190744400024414,
|
| 288 |
+
"learning_rate": 1.714530524812567e-05,
|
| 289 |
+
"loss": 1.6801,
|
| 290 |
+
"step": 2000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.7318814709032488,
|
| 294 |
+
"grad_norm": 14.35610580444336,
|
| 295 |
+
"learning_rate": 1.7073902177793647e-05,
|
| 296 |
+
"loss": 1.6561,
|
| 297 |
+
"step": 2050
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.7497322384862549,
|
| 301 |
+
"grad_norm": 18.017112731933594,
|
| 302 |
+
"learning_rate": 1.7002499107461622e-05,
|
| 303 |
+
"loss": 1.6541,
|
| 304 |
+
"step": 2100
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.767583006069261,
|
| 308 |
+
"grad_norm": 17.52274513244629,
|
| 309 |
+
"learning_rate": 1.6931096037129598e-05,
|
| 310 |
+
"loss": 1.6398,
|
| 311 |
+
"step": 2150
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.7854337736522671,
|
| 315 |
+
"grad_norm": 10.50312328338623,
|
| 316 |
+
"learning_rate": 1.6859692966797574e-05,
|
| 317 |
+
"loss": 1.5567,
|
| 318 |
+
"step": 2200
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.8032845412352732,
|
| 322 |
+
"grad_norm": 13.701729774475098,
|
| 323 |
+
"learning_rate": 1.678828989646555e-05,
|
| 324 |
+
"loss": 1.5603,
|
| 325 |
+
"step": 2250
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.8211353088182792,
|
| 329 |
+
"grad_norm": 11.937254905700684,
|
| 330 |
+
"learning_rate": 1.6716886826133525e-05,
|
| 331 |
+
"loss": 1.6716,
|
| 332 |
+
"step": 2300
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.8389860764012853,
|
| 336 |
+
"grad_norm": 17.40386199951172,
|
| 337 |
+
"learning_rate": 1.66454837558015e-05,
|
| 338 |
+
"loss": 1.6773,
|
| 339 |
+
"step": 2350
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.8568368439842913,
|
| 343 |
+
"grad_norm": 15.239500045776367,
|
| 344 |
+
"learning_rate": 1.6574080685469476e-05,
|
| 345 |
+
"loss": 1.596,
|
| 346 |
+
"step": 2400
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.8746876115672974,
|
| 350 |
+
"grad_norm": 14.679793357849121,
|
| 351 |
+
"learning_rate": 1.650267761513745e-05,
|
| 352 |
+
"loss": 1.5855,
|
| 353 |
+
"step": 2450
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.8925383791503034,
|
| 357 |
+
"grad_norm": 12.250946998596191,
|
| 358 |
+
"learning_rate": 1.643127454480543e-05,
|
| 359 |
+
"loss": 1.6029,
|
| 360 |
+
"step": 2500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.9103891467333095,
|
| 364 |
+
"grad_norm": 14.957499504089355,
|
| 365 |
+
"learning_rate": 1.6359871474473403e-05,
|
| 366 |
+
"loss": 1.6466,
|
| 367 |
+
"step": 2550
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.9282399143163156,
|
| 371 |
+
"grad_norm": 21.005565643310547,
|
| 372 |
+
"learning_rate": 1.628846840414138e-05,
|
| 373 |
+
"loss": 1.5891,
|
| 374 |
+
"step": 2600
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.9460906818993217,
|
| 378 |
+
"grad_norm": 13.731078147888184,
|
| 379 |
+
"learning_rate": 1.6217065333809354e-05,
|
| 380 |
+
"loss": 1.5669,
|
| 381 |
+
"step": 2650
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.9639414494823277,
|
| 385 |
+
"grad_norm": 14.127930641174316,
|
| 386 |
+
"learning_rate": 1.614566226347733e-05,
|
| 387 |
+
"loss": 1.5489,
|
| 388 |
+
"step": 2700
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.9817922170653338,
|
| 392 |
+
"grad_norm": 12.521357536315918,
|
| 393 |
+
"learning_rate": 1.607425919314531e-05,
|
| 394 |
+
"loss": 1.5861,
|
| 395 |
+
"step": 2750
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.9996429846483399,
|
| 399 |
+
"grad_norm": 12.26593017578125,
|
| 400 |
+
"learning_rate": 1.600285612281328e-05,
|
| 401 |
+
"loss": 1.4877,
|
| 402 |
+
"step": 2800
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.0,
|
| 406 |
+
"eval_accuracy": 0.47514058734267606,
|
| 407 |
+
"eval_loss": 1.4969408512115479,
|
| 408 |
+
"eval_macro_f1": 0.4136955039973382,
|
| 409 |
+
"eval_runtime": 5.1653,
|
| 410 |
+
"eval_samples_per_second": 2168.895,
|
| 411 |
+
"eval_steps_per_second": 34.074,
|
| 412 |
+
"step": 2801
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.0174937522313459,
|
| 416 |
+
"grad_norm": 15.36621379852295,
|
| 417 |
+
"learning_rate": 1.593145305248126e-05,
|
| 418 |
+
"loss": 1.4006,
|
| 419 |
+
"step": 2850
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.035344519814352,
|
| 423 |
+
"grad_norm": 12.056408882141113,
|
| 424 |
+
"learning_rate": 1.5860049982149232e-05,
|
| 425 |
+
"loss": 1.3621,
|
| 426 |
+
"step": 2900
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.053195287397358,
|
| 430 |
+
"grad_norm": 10.554118156433105,
|
| 431 |
+
"learning_rate": 1.5788646911817208e-05,
|
| 432 |
+
"loss": 1.3599,
|
| 433 |
+
"step": 2950
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 1.0710460549803642,
|
| 437 |
+
"grad_norm": 13.228151321411133,
|
| 438 |
+
"learning_rate": 1.5717243841485187e-05,
|
| 439 |
+
"loss": 1.4433,
|
| 440 |
+
"step": 3000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 1.0888968225633702,
|
| 444 |
+
"grad_norm": 19.01808738708496,
|
| 445 |
+
"learning_rate": 1.564584077115316e-05,
|
| 446 |
+
"loss": 1.3759,
|
| 447 |
+
"step": 3050
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.1067475901463764,
|
| 451 |
+
"grad_norm": 12.861035346984863,
|
| 452 |
+
"learning_rate": 1.5574437700821138e-05,
|
| 453 |
+
"loss": 1.4442,
|
| 454 |
+
"step": 3100
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 1.1245983577293823,
|
| 458 |
+
"grad_norm": 9.24783706665039,
|
| 459 |
+
"learning_rate": 1.5503034630489113e-05,
|
| 460 |
+
"loss": 1.3682,
|
| 461 |
+
"step": 3150
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.1424491253123885,
|
| 465 |
+
"grad_norm": 15.15107250213623,
|
| 466 |
+
"learning_rate": 1.543163156015709e-05,
|
| 467 |
+
"loss": 1.3637,
|
| 468 |
+
"step": 3200
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 1.1602998928953945,
|
| 472 |
+
"grad_norm": 12.833767890930176,
|
| 473 |
+
"learning_rate": 1.5360228489825065e-05,
|
| 474 |
+
"loss": 1.3372,
|
| 475 |
+
"step": 3250
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 1.1781506604784004,
|
| 479 |
+
"grad_norm": 12.872199058532715,
|
| 480 |
+
"learning_rate": 1.528882541949304e-05,
|
| 481 |
+
"loss": 1.4134,
|
| 482 |
+
"step": 3300
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.1960014280614066,
|
| 486 |
+
"grad_norm": 12.539731979370117,
|
| 487 |
+
"learning_rate": 1.5217422349161016e-05,
|
| 488 |
+
"loss": 1.3883,
|
| 489 |
+
"step": 3350
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.2138521956444128,
|
| 493 |
+
"grad_norm": 14.508119583129883,
|
| 494 |
+
"learning_rate": 1.5146019278828991e-05,
|
| 495 |
+
"loss": 1.3427,
|
| 496 |
+
"step": 3400
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.2317029632274188,
|
| 500 |
+
"grad_norm": 18.304662704467773,
|
| 501 |
+
"learning_rate": 1.5074616208496965e-05,
|
| 502 |
+
"loss": 1.3978,
|
| 503 |
+
"step": 3450
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.2495537308104248,
|
| 507 |
+
"grad_norm": 15.10910701751709,
|
| 508 |
+
"learning_rate": 1.5003213138164943e-05,
|
| 509 |
+
"loss": 1.389,
|
| 510 |
+
"step": 3500
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.267404498393431,
|
| 514 |
+
"grad_norm": 12.175374984741211,
|
| 515 |
+
"learning_rate": 1.4931810067832916e-05,
|
| 516 |
+
"loss": 1.2985,
|
| 517 |
+
"step": 3550
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 1.285255265976437,
|
| 521 |
+
"grad_norm": 11.14751148223877,
|
| 522 |
+
"learning_rate": 1.4860406997500894e-05,
|
| 523 |
+
"loss": 1.384,
|
| 524 |
+
"step": 3600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.303106033559443,
|
| 528 |
+
"grad_norm": 15.333986282348633,
|
| 529 |
+
"learning_rate": 1.478900392716887e-05,
|
| 530 |
+
"loss": 1.2794,
|
| 531 |
+
"step": 3650
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.320956801142449,
|
| 535 |
+
"grad_norm": 15.89001178741455,
|
| 536 |
+
"learning_rate": 1.4717600856836845e-05,
|
| 537 |
+
"loss": 1.2597,
|
| 538 |
+
"step": 3700
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 1.3388075687254553,
|
| 542 |
+
"grad_norm": 9.596504211425781,
|
| 543 |
+
"learning_rate": 1.464619778650482e-05,
|
| 544 |
+
"loss": 1.358,
|
| 545 |
+
"step": 3750
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.3566583363084612,
|
| 549 |
+
"grad_norm": 12.051373481750488,
|
| 550 |
+
"learning_rate": 1.4574794716172798e-05,
|
| 551 |
+
"loss": 1.3429,
|
| 552 |
+
"step": 3800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.3745091038914674,
|
| 556 |
+
"grad_norm": 9.407623291015625,
|
| 557 |
+
"learning_rate": 1.4503391645840772e-05,
|
| 558 |
+
"loss": 1.3347,
|
| 559 |
+
"step": 3850
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.3923598714744734,
|
| 563 |
+
"grad_norm": 13.37045669555664,
|
| 564 |
+
"learning_rate": 1.4431988575508749e-05,
|
| 565 |
+
"loss": 1.3214,
|
| 566 |
+
"step": 3900
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.4102106390574796,
|
| 570 |
+
"grad_norm": 14.173185348510742,
|
| 571 |
+
"learning_rate": 1.4360585505176723e-05,
|
| 572 |
+
"loss": 1.3509,
|
| 573 |
+
"step": 3950
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.4280614066404855,
|
| 577 |
+
"grad_norm": 20.702144622802734,
|
| 578 |
+
"learning_rate": 1.42891824348447e-05,
|
| 579 |
+
"loss": 1.3057,
|
| 580 |
+
"step": 4000
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.4459121742234915,
|
| 584 |
+
"grad_norm": 11.873905181884766,
|
| 585 |
+
"learning_rate": 1.4217779364512676e-05,
|
| 586 |
+
"loss": 1.333,
|
| 587 |
+
"step": 4050
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.4637629418064977,
|
| 591 |
+
"grad_norm": 12.91700553894043,
|
| 592 |
+
"learning_rate": 1.414637629418065e-05,
|
| 593 |
+
"loss": 1.2759,
|
| 594 |
+
"step": 4100
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.4816137093895039,
|
| 598 |
+
"grad_norm": 16.117185592651367,
|
| 599 |
+
"learning_rate": 1.4074973223848627e-05,
|
| 600 |
+
"loss": 1.2325,
|
| 601 |
+
"step": 4150
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.4994644769725098,
|
| 605 |
+
"grad_norm": 12.514154434204102,
|
| 606 |
+
"learning_rate": 1.40035701535166e-05,
|
| 607 |
+
"loss": 1.3223,
|
| 608 |
+
"step": 4200
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.5173152445555158,
|
| 612 |
+
"grad_norm": 19.87337303161621,
|
| 613 |
+
"learning_rate": 1.3932167083184578e-05,
|
| 614 |
+
"loss": 1.3427,
|
| 615 |
+
"step": 4250
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.535166012138522,
|
| 619 |
+
"grad_norm": 13.149452209472656,
|
| 620 |
+
"learning_rate": 1.3860764012852554e-05,
|
| 621 |
+
"loss": 1.3193,
|
| 622 |
+
"step": 4300
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 1.5530167797215282,
|
| 626 |
+
"grad_norm": 16.208499908447266,
|
| 627 |
+
"learning_rate": 1.378936094252053e-05,
|
| 628 |
+
"loss": 1.3719,
|
| 629 |
+
"step": 4350
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 1.570867547304534,
|
| 633 |
+
"grad_norm": 15.929703712463379,
|
| 634 |
+
"learning_rate": 1.3717957872188505e-05,
|
| 635 |
+
"loss": 1.369,
|
| 636 |
+
"step": 4400
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 1.5887183148875401,
|
| 640 |
+
"grad_norm": 12.900412559509277,
|
| 641 |
+
"learning_rate": 1.3646554801856482e-05,
|
| 642 |
+
"loss": 1.3369,
|
| 643 |
+
"step": 4450
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.6065690824705463,
|
| 647 |
+
"grad_norm": 13.354098320007324,
|
| 648 |
+
"learning_rate": 1.3575151731524456e-05,
|
| 649 |
+
"loss": 1.2681,
|
| 650 |
+
"step": 4500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.6244198500535523,
|
| 654 |
+
"grad_norm": 17.960058212280273,
|
| 655 |
+
"learning_rate": 1.3503748661192433e-05,
|
| 656 |
+
"loss": 1.2589,
|
| 657 |
+
"step": 4550
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.6422706176365582,
|
| 661 |
+
"grad_norm": 17.655719757080078,
|
| 662 |
+
"learning_rate": 1.3432345590860407e-05,
|
| 663 |
+
"loss": 1.3318,
|
| 664 |
+
"step": 4600
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 1.6601213852195644,
|
| 668 |
+
"grad_norm": 19.73931121826172,
|
| 669 |
+
"learning_rate": 1.3360942520528385e-05,
|
| 670 |
+
"loss": 1.2895,
|
| 671 |
+
"step": 4650
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 1.6779721528025706,
|
| 675 |
+
"grad_norm": 14.323142051696777,
|
| 676 |
+
"learning_rate": 1.328953945019636e-05,
|
| 677 |
+
"loss": 1.2691,
|
| 678 |
+
"step": 4700
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 1.6958229203855766,
|
| 682 |
+
"grad_norm": 8.690425872802734,
|
| 683 |
+
"learning_rate": 1.3218136379864334e-05,
|
| 684 |
+
"loss": 1.2606,
|
| 685 |
+
"step": 4750
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.7136736879685825,
|
| 689 |
+
"grad_norm": 10.671348571777344,
|
| 690 |
+
"learning_rate": 1.3146733309532311e-05,
|
| 691 |
+
"loss": 1.3214,
|
| 692 |
+
"step": 4800
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.7315244555515887,
|
| 696 |
+
"grad_norm": 8.343477249145508,
|
| 697 |
+
"learning_rate": 1.3075330239200285e-05,
|
| 698 |
+
"loss": 1.2785,
|
| 699 |
+
"step": 4850
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.749375223134595,
|
| 703 |
+
"grad_norm": 22.884681701660156,
|
| 704 |
+
"learning_rate": 1.3003927168868263e-05,
|
| 705 |
+
"loss": 1.2526,
|
| 706 |
+
"step": 4900
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 1.767225990717601,
|
| 710 |
+
"grad_norm": 17.500192642211914,
|
| 711 |
+
"learning_rate": 1.2932524098536238e-05,
|
| 712 |
+
"loss": 1.2738,
|
| 713 |
+
"step": 4950
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 1.7850767583006069,
|
| 717 |
+
"grad_norm": 17.966121673583984,
|
| 718 |
+
"learning_rate": 1.2861121028204214e-05,
|
| 719 |
+
"loss": 1.2983,
|
| 720 |
+
"step": 5000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 1.802927525883613,
|
| 724 |
+
"grad_norm": 22.048490524291992,
|
| 725 |
+
"learning_rate": 1.278971795787219e-05,
|
| 726 |
+
"loss": 1.3437,
|
| 727 |
+
"step": 5050
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.820778293466619,
|
| 731 |
+
"grad_norm": 16.436044692993164,
|
| 732 |
+
"learning_rate": 1.2718314887540167e-05,
|
| 733 |
+
"loss": 1.2604,
|
| 734 |
+
"step": 5100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.838629061049625,
|
| 738 |
+
"grad_norm": 12.748188972473145,
|
| 739 |
+
"learning_rate": 1.264691181720814e-05,
|
| 740 |
+
"loss": 1.1955,
|
| 741 |
+
"step": 5150
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 1.8564798286326312,
|
| 745 |
+
"grad_norm": 15.320619583129883,
|
| 746 |
+
"learning_rate": 1.2575508746876118e-05,
|
| 747 |
+
"loss": 1.2902,
|
| 748 |
+
"step": 5200
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.8743305962156374,
|
| 752 |
+
"grad_norm": 17.871013641357422,
|
| 753 |
+
"learning_rate": 1.2504105676544092e-05,
|
| 754 |
+
"loss": 1.3103,
|
| 755 |
+
"step": 5250
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.8921813637986433,
|
| 759 |
+
"grad_norm": 16.096118927001953,
|
| 760 |
+
"learning_rate": 1.2432702606212067e-05,
|
| 761 |
+
"loss": 1.3319,
|
| 762 |
+
"step": 5300
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.9100321313816493,
|
| 766 |
+
"grad_norm": 15.915545463562012,
|
| 767 |
+
"learning_rate": 1.2361299535880045e-05,
|
| 768 |
+
"loss": 1.3093,
|
| 769 |
+
"step": 5350
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.9278828989646555,
|
| 773 |
+
"grad_norm": 12.922966003417969,
|
| 774 |
+
"learning_rate": 1.2289896465548019e-05,
|
| 775 |
+
"loss": 1.2533,
|
| 776 |
+
"step": 5400
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.9457336665476617,
|
| 780 |
+
"grad_norm": 20.724853515625,
|
| 781 |
+
"learning_rate": 1.2218493395215996e-05,
|
| 782 |
+
"loss": 1.2531,
|
| 783 |
+
"step": 5450
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 1.9635844341306676,
|
| 787 |
+
"grad_norm": 13.744089126586914,
|
| 788 |
+
"learning_rate": 1.214709032488397e-05,
|
| 789 |
+
"loss": 1.2476,
|
| 790 |
+
"step": 5500
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.9814352017136736,
|
| 794 |
+
"grad_norm": 11.663731575012207,
|
| 795 |
+
"learning_rate": 1.2075687254551947e-05,
|
| 796 |
+
"loss": 1.2768,
|
| 797 |
+
"step": 5550
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 1.9992859692966798,
|
| 801 |
+
"grad_norm": 16.515056610107422,
|
| 802 |
+
"learning_rate": 1.2004284184219923e-05,
|
| 803 |
+
"loss": 1.283,
|
| 804 |
+
"step": 5600
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 2.0,
|
| 808 |
+
"eval_accuracy": 0.5697581005087923,
|
| 809 |
+
"eval_loss": 1.3204034566879272,
|
| 810 |
+
"eval_macro_f1": 0.48569754853380576,
|
| 811 |
+
"eval_runtime": 5.4967,
|
| 812 |
+
"eval_samples_per_second": 2038.126,
|
| 813 |
+
"eval_steps_per_second": 32.019,
|
| 814 |
+
"step": 5602
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 2.017136736879686,
|
| 818 |
+
"grad_norm": 10.937089920043945,
|
| 819 |
+
"learning_rate": 1.1932881113887898e-05,
|
| 820 |
+
"loss": 1.0781,
|
| 821 |
+
"step": 5650
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 2.0349875044626917,
|
| 825 |
+
"grad_norm": 17.1431827545166,
|
| 826 |
+
"learning_rate": 1.1861478043555874e-05,
|
| 827 |
+
"loss": 1.034,
|
| 828 |
+
"step": 5700
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 2.052838272045698,
|
| 832 |
+
"grad_norm": 12.69800853729248,
|
| 833 |
+
"learning_rate": 1.1790074973223851e-05,
|
| 834 |
+
"loss": 1.0675,
|
| 835 |
+
"step": 5750
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 2.070689039628704,
|
| 839 |
+
"grad_norm": 19.09876251220703,
|
| 840 |
+
"learning_rate": 1.1718671902891825e-05,
|
| 841 |
+
"loss": 1.0352,
|
| 842 |
+
"step": 5800
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 2.0885398072117103,
|
| 846 |
+
"grad_norm": 16.88023567199707,
|
| 847 |
+
"learning_rate": 1.1647268832559802e-05,
|
| 848 |
+
"loss": 0.9982,
|
| 849 |
+
"step": 5850
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.106390574794716,
|
| 853 |
+
"grad_norm": 9.332680702209473,
|
| 854 |
+
"learning_rate": 1.1575865762227776e-05,
|
| 855 |
+
"loss": 1.058,
|
| 856 |
+
"step": 5900
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 2.124241342377722,
|
| 860 |
+
"grad_norm": 15.302008628845215,
|
| 861 |
+
"learning_rate": 1.1504462691895752e-05,
|
| 862 |
+
"loss": 1.0775,
|
| 863 |
+
"step": 5950
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 2.1420921099607284,
|
| 867 |
+
"grad_norm": 13.956599235534668,
|
| 868 |
+
"learning_rate": 1.1433059621563729e-05,
|
| 869 |
+
"loss": 1.0847,
|
| 870 |
+
"step": 6000
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 2.1599428775437346,
|
| 874 |
+
"grad_norm": 13.09730052947998,
|
| 875 |
+
"learning_rate": 1.1361656551231703e-05,
|
| 876 |
+
"loss": 1.0621,
|
| 877 |
+
"step": 6050
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 2.1777936451267403,
|
| 881 |
+
"grad_norm": 18.93950080871582,
|
| 882 |
+
"learning_rate": 1.129025348089968e-05,
|
| 883 |
+
"loss": 1.0375,
|
| 884 |
+
"step": 6100
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.1956444127097465,
|
| 888 |
+
"grad_norm": 16.003028869628906,
|
| 889 |
+
"learning_rate": 1.1218850410567654e-05,
|
| 890 |
+
"loss": 1.0133,
|
| 891 |
+
"step": 6150
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.2134951802927527,
|
| 895 |
+
"grad_norm": 11.415363311767578,
|
| 896 |
+
"learning_rate": 1.1147447340235631e-05,
|
| 897 |
+
"loss": 1.0313,
|
| 898 |
+
"step": 6200
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 2.2313459478757585,
|
| 902 |
+
"grad_norm": 15.419634819030762,
|
| 903 |
+
"learning_rate": 1.1076044269903607e-05,
|
| 904 |
+
"loss": 1.0583,
|
| 905 |
+
"step": 6250
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.2491967154587647,
|
| 909 |
+
"grad_norm": 16.330265045166016,
|
| 910 |
+
"learning_rate": 1.1004641199571583e-05,
|
| 911 |
+
"loss": 0.9851,
|
| 912 |
+
"step": 6300
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 2.267047483041771,
|
| 916 |
+
"grad_norm": 12.965764999389648,
|
| 917 |
+
"learning_rate": 1.0933238129239558e-05,
|
| 918 |
+
"loss": 1.0314,
|
| 919 |
+
"step": 6350
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 2.284898250624777,
|
| 923 |
+
"grad_norm": 13.175607681274414,
|
| 924 |
+
"learning_rate": 1.0861835058907535e-05,
|
| 925 |
+
"loss": 1.0458,
|
| 926 |
+
"step": 6400
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.3027490182077828,
|
| 930 |
+
"grad_norm": 17.50844383239746,
|
| 931 |
+
"learning_rate": 1.079043198857551e-05,
|
| 932 |
+
"loss": 1.0705,
|
| 933 |
+
"step": 6450
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.320599785790789,
|
| 937 |
+
"grad_norm": 11.109663009643555,
|
| 938 |
+
"learning_rate": 1.0719028918243487e-05,
|
| 939 |
+
"loss": 1.1261,
|
| 940 |
+
"step": 6500
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 2.338450553373795,
|
| 944 |
+
"grad_norm": 22.36432456970215,
|
| 945 |
+
"learning_rate": 1.064762584791146e-05,
|
| 946 |
+
"loss": 1.0871,
|
| 947 |
+
"step": 6550
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 2.356301320956801,
|
| 951 |
+
"grad_norm": 12.304158210754395,
|
| 952 |
+
"learning_rate": 1.0576222777579436e-05,
|
| 953 |
+
"loss": 1.0667,
|
| 954 |
+
"step": 6600
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 2.374152088539807,
|
| 958 |
+
"grad_norm": 17.145753860473633,
|
| 959 |
+
"learning_rate": 1.0504819707247413e-05,
|
| 960 |
+
"loss": 1.0315,
|
| 961 |
+
"step": 6650
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 2.3920028561228133,
|
| 965 |
+
"grad_norm": 10.1165189743042,
|
| 966 |
+
"learning_rate": 1.0433416636915387e-05,
|
| 967 |
+
"loss": 1.0643,
|
| 968 |
+
"step": 6700
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.4098536237058195,
|
| 972 |
+
"grad_norm": 11.35006332397461,
|
| 973 |
+
"learning_rate": 1.0362013566583365e-05,
|
| 974 |
+
"loss": 1.0621,
|
| 975 |
+
"step": 6750
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 2.4277043912888256,
|
| 979 |
+
"grad_norm": 13.226170539855957,
|
| 980 |
+
"learning_rate": 1.0290610496251339e-05,
|
| 981 |
+
"loss": 1.0086,
|
| 982 |
+
"step": 6800
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 2.4455551588718314,
|
| 986 |
+
"grad_norm": 13.296606063842773,
|
| 987 |
+
"learning_rate": 1.0219207425919316e-05,
|
| 988 |
+
"loss": 1.0329,
|
| 989 |
+
"step": 6850
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 2.4634059264548376,
|
| 993 |
+
"grad_norm": 11.565988540649414,
|
| 994 |
+
"learning_rate": 1.0147804355587291e-05,
|
| 995 |
+
"loss": 1.0779,
|
| 996 |
+
"step": 6900
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 2.4812566940378438,
|
| 1000 |
+
"grad_norm": 21.814311981201172,
|
| 1001 |
+
"learning_rate": 1.0076401285255265e-05,
|
| 1002 |
+
"loss": 1.0901,
|
| 1003 |
+
"step": 6950
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 2.4991074616208495,
|
| 1007 |
+
"grad_norm": 20.675832748413086,
|
| 1008 |
+
"learning_rate": 1.0004998214923243e-05,
|
| 1009 |
+
"loss": 1.0447,
|
| 1010 |
+
"step": 7000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.5169582292038557,
|
| 1014 |
+
"grad_norm": 18.146474838256836,
|
| 1015 |
+
"learning_rate": 9.933595144591218e-06,
|
| 1016 |
+
"loss": 1.1045,
|
| 1017 |
+
"step": 7050
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.534808996786862,
|
| 1021 |
+
"grad_norm": 14.326371192932129,
|
| 1022 |
+
"learning_rate": 9.862192074259194e-06,
|
| 1023 |
+
"loss": 1.0594,
|
| 1024 |
+
"step": 7100
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 2.552659764369868,
|
| 1028 |
+
"grad_norm": 14.145441055297852,
|
| 1029 |
+
"learning_rate": 9.79078900392717e-06,
|
| 1030 |
+
"loss": 1.0402,
|
| 1031 |
+
"step": 7150
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 2.570510531952874,
|
| 1035 |
+
"grad_norm": 17.572402954101562,
|
| 1036 |
+
"learning_rate": 9.719385933595145e-06,
|
| 1037 |
+
"loss": 1.0297,
|
| 1038 |
+
"step": 7200
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 2.58836129953588,
|
| 1042 |
+
"grad_norm": 16.921300888061523,
|
| 1043 |
+
"learning_rate": 9.64798286326312e-06,
|
| 1044 |
+
"loss": 1.0694,
|
| 1045 |
+
"step": 7250
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 2.606212067118886,
|
| 1049 |
+
"grad_norm": 9.290846824645996,
|
| 1050 |
+
"learning_rate": 9.576579792931096e-06,
|
| 1051 |
+
"loss": 1.0014,
|
| 1052 |
+
"step": 7300
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.624062834701892,
|
| 1056 |
+
"grad_norm": 18.461746215820312,
|
| 1057 |
+
"learning_rate": 9.505176722599072e-06,
|
| 1058 |
+
"loss": 1.0958,
|
| 1059 |
+
"step": 7350
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 2.641913602284898,
|
| 1063 |
+
"grad_norm": 18.638286590576172,
|
| 1064 |
+
"learning_rate": 9.433773652267049e-06,
|
| 1065 |
+
"loss": 1.0575,
|
| 1066 |
+
"step": 7400
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 2.6597643698679043,
|
| 1070 |
+
"grad_norm": 20.554609298706055,
|
| 1071 |
+
"learning_rate": 9.362370581935025e-06,
|
| 1072 |
+
"loss": 1.0293,
|
| 1073 |
+
"step": 7450
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.6776151374509105,
|
| 1077 |
+
"grad_norm": 17.582111358642578,
|
| 1078 |
+
"learning_rate": 9.290967511603e-06,
|
| 1079 |
+
"loss": 1.0127,
|
| 1080 |
+
"step": 7500
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.6954659050339167,
|
| 1084 |
+
"grad_norm": 18.36084747314453,
|
| 1085 |
+
"learning_rate": 9.219564441270976e-06,
|
| 1086 |
+
"loss": 1.0402,
|
| 1087 |
+
"step": 7550
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 2.7133166726169224,
|
| 1091 |
+
"grad_norm": 11.184069633483887,
|
| 1092 |
+
"learning_rate": 9.148161370938951e-06,
|
| 1093 |
+
"loss": 1.0198,
|
| 1094 |
+
"step": 7600
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.7311674401999286,
|
| 1098 |
+
"grad_norm": 19.872299194335938,
|
| 1099 |
+
"learning_rate": 9.076758300606927e-06,
|
| 1100 |
+
"loss": 1.0562,
|
| 1101 |
+
"step": 7650
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 2.749018207782935,
|
| 1105 |
+
"grad_norm": 18.872329711914062,
|
| 1106 |
+
"learning_rate": 9.005355230274903e-06,
|
| 1107 |
+
"loss": 1.0907,
|
| 1108 |
+
"step": 7700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 2.7668689753659406,
|
| 1112 |
+
"grad_norm": 21.884220123291016,
|
| 1113 |
+
"learning_rate": 8.933952159942878e-06,
|
| 1114 |
+
"loss": 1.0713,
|
| 1115 |
+
"step": 7750
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.7847197429489468,
|
| 1119 |
+
"grad_norm": 20.70819854736328,
|
| 1120 |
+
"learning_rate": 8.862549089610854e-06,
|
| 1121 |
+
"loss": 1.001,
|
| 1122 |
+
"step": 7800
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 2.802570510531953,
|
| 1126 |
+
"grad_norm": 33.434959411621094,
|
| 1127 |
+
"learning_rate": 8.79114601927883e-06,
|
| 1128 |
+
"loss": 1.0503,
|
| 1129 |
+
"step": 7850
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 2.820421278114959,
|
| 1133 |
+
"grad_norm": 16.19710922241211,
|
| 1134 |
+
"learning_rate": 8.719742948946805e-06,
|
| 1135 |
+
"loss": 1.0701,
|
| 1136 |
+
"step": 7900
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.838272045697965,
|
| 1140 |
+
"grad_norm": 14.586867332458496,
|
| 1141 |
+
"learning_rate": 8.64833987861478e-06,
|
| 1142 |
+
"loss": 1.0619,
|
| 1143 |
+
"step": 7950
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 2.856122813280971,
|
| 1147 |
+
"grad_norm": 15.705036163330078,
|
| 1148 |
+
"learning_rate": 8.576936808282756e-06,
|
| 1149 |
+
"loss": 1.0708,
|
| 1150 |
+
"step": 8000
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 2.8739735808639773,
|
| 1154 |
+
"grad_norm": 14.276032447814941,
|
| 1155 |
+
"learning_rate": 8.505533737950734e-06,
|
| 1156 |
+
"loss": 1.1002,
|
| 1157 |
+
"step": 8050
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.891824348446983,
|
| 1161 |
+
"grad_norm": 19.014434814453125,
|
| 1162 |
+
"learning_rate": 8.434130667618709e-06,
|
| 1163 |
+
"loss": 0.9907,
|
| 1164 |
+
"step": 8100
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 2.909675116029989,
|
| 1168 |
+
"grad_norm": 20.64139747619629,
|
| 1169 |
+
"learning_rate": 8.362727597286685e-06,
|
| 1170 |
+
"loss": 1.0514,
|
| 1171 |
+
"step": 8150
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 2.9275258836129954,
|
| 1175 |
+
"grad_norm": 19.546451568603516,
|
| 1176 |
+
"learning_rate": 8.291324526954659e-06,
|
| 1177 |
+
"loss": 1.059,
|
| 1178 |
+
"step": 8200
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.9453766511960016,
|
| 1182 |
+
"grad_norm": 18.8275203704834,
|
| 1183 |
+
"learning_rate": 8.219921456622636e-06,
|
| 1184 |
+
"loss": 1.0642,
|
| 1185 |
+
"step": 8250
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 2.9632274187790077,
|
| 1189 |
+
"grad_norm": 14.960760116577148,
|
| 1190 |
+
"learning_rate": 8.148518386290611e-06,
|
| 1191 |
+
"loss": 1.1098,
|
| 1192 |
+
"step": 8300
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 2.9810781863620135,
|
| 1196 |
+
"grad_norm": 11.512022972106934,
|
| 1197 |
+
"learning_rate": 8.077115315958587e-06,
|
| 1198 |
+
"loss": 1.0295,
|
| 1199 |
+
"step": 8350
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.9989289539450197,
|
| 1203 |
+
"grad_norm": 18.68143653869629,
|
| 1204 |
+
"learning_rate": 8.005712245626563e-06,
|
| 1205 |
+
"loss": 1.0402,
|
| 1206 |
+
"step": 8400
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 3.0,
|
| 1210 |
+
"eval_accuracy": 0.5899312684102472,
|
| 1211 |
+
"eval_loss": 1.2934809923171997,
|
| 1212 |
+
"eval_macro_f1": 0.5093257341608811,
|
| 1213 |
+
"eval_runtime": 5.2479,
|
| 1214 |
+
"eval_samples_per_second": 2134.777,
|
| 1215 |
+
"eval_steps_per_second": 33.538,
|
| 1216 |
+
"step": 8403
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 3.016779721528026,
|
| 1220 |
+
"grad_norm": 19.454416275024414,
|
| 1221 |
+
"learning_rate": 7.934309175294538e-06,
|
| 1222 |
+
"loss": 0.8671,
|
| 1223 |
+
"step": 8450
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 3.0346304891110316,
|
| 1227 |
+
"grad_norm": 14.609670639038086,
|
| 1228 |
+
"learning_rate": 7.862906104962514e-06,
|
| 1229 |
+
"loss": 0.8195,
|
| 1230 |
+
"step": 8500
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 3.052481256694038,
|
| 1234 |
+
"grad_norm": 13.611706733703613,
|
| 1235 |
+
"learning_rate": 7.79150303463049e-06,
|
| 1236 |
+
"loss": 0.9039,
|
| 1237 |
+
"step": 8550
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 3.070332024277044,
|
| 1241 |
+
"grad_norm": 17.289657592773438,
|
| 1242 |
+
"learning_rate": 7.720099964298465e-06,
|
| 1243 |
+
"loss": 0.8817,
|
| 1244 |
+
"step": 8600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 3.08818279186005,
|
| 1248 |
+
"grad_norm": 12.592385292053223,
|
| 1249 |
+
"learning_rate": 7.64869689396644e-06,
|
| 1250 |
+
"loss": 0.8339,
|
| 1251 |
+
"step": 8650
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 3.106033559443056,
|
| 1255 |
+
"grad_norm": 17.638628005981445,
|
| 1256 |
+
"learning_rate": 7.577293823634417e-06,
|
| 1257 |
+
"loss": 0.8151,
|
| 1258 |
+
"step": 8700
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 3.123884327026062,
|
| 1262 |
+
"grad_norm": 11.680763244628906,
|
| 1263 |
+
"learning_rate": 7.505890753302393e-06,
|
| 1264 |
+
"loss": 0.8438,
|
| 1265 |
+
"step": 8750
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 3.1417350946090683,
|
| 1269 |
+
"grad_norm": 22.238727569580078,
|
| 1270 |
+
"learning_rate": 7.434487682970368e-06,
|
| 1271 |
+
"loss": 0.8263,
|
| 1272 |
+
"step": 8800
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 3.159585862192074,
|
| 1276 |
+
"grad_norm": 13.042167663574219,
|
| 1277 |
+
"learning_rate": 7.363084612638344e-06,
|
| 1278 |
+
"loss": 0.8538,
|
| 1279 |
+
"step": 8850
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 3.1774366297750802,
|
| 1283 |
+
"grad_norm": 11.398730278015137,
|
| 1284 |
+
"learning_rate": 7.29168154230632e-06,
|
| 1285 |
+
"loss": 0.8465,
|
| 1286 |
+
"step": 8900
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 3.1952873973580864,
|
| 1290 |
+
"grad_norm": 12.321819305419922,
|
| 1291 |
+
"learning_rate": 7.220278471974296e-06,
|
| 1292 |
+
"loss": 0.8297,
|
| 1293 |
+
"step": 8950
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 3.2131381649410926,
|
| 1297 |
+
"grad_norm": 10.085368156433105,
|
| 1298 |
+
"learning_rate": 7.1488754016422715e-06,
|
| 1299 |
+
"loss": 0.8655,
|
| 1300 |
+
"step": 9000
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 3.2309889325240984,
|
| 1304 |
+
"grad_norm": 7.707581043243408,
|
| 1305 |
+
"learning_rate": 7.077472331310247e-06,
|
| 1306 |
+
"loss": 0.7967,
|
| 1307 |
+
"step": 9050
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 3.2488397001071045,
|
| 1311 |
+
"grad_norm": 13.6712646484375,
|
| 1312 |
+
"learning_rate": 7.006069260978222e-06,
|
| 1313 |
+
"loss": 0.8858,
|
| 1314 |
+
"step": 9100
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 3.2666904676901107,
|
| 1318 |
+
"grad_norm": 24.826183319091797,
|
| 1319 |
+
"learning_rate": 6.934666190646199e-06,
|
| 1320 |
+
"loss": 0.8799,
|
| 1321 |
+
"step": 9150
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 3.284541235273117,
|
| 1325 |
+
"grad_norm": 16.819063186645508,
|
| 1326 |
+
"learning_rate": 6.863263120314174e-06,
|
| 1327 |
+
"loss": 0.889,
|
| 1328 |
+
"step": 9200
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 3.3023920028561227,
|
| 1332 |
+
"grad_norm": 16.014638900756836,
|
| 1333 |
+
"learning_rate": 6.7918600499821495e-06,
|
| 1334 |
+
"loss": 0.8766,
|
| 1335 |
+
"step": 9250
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 3.320242770439129,
|
| 1339 |
+
"grad_norm": 16.722623825073242,
|
| 1340 |
+
"learning_rate": 6.720456979650125e-06,
|
| 1341 |
+
"loss": 0.8391,
|
| 1342 |
+
"step": 9300
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 3.338093538022135,
|
| 1346 |
+
"grad_norm": 15.532989501953125,
|
| 1347 |
+
"learning_rate": 6.6490539093181015e-06,
|
| 1348 |
+
"loss": 0.906,
|
| 1349 |
+
"step": 9350
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 3.3559443056051412,
|
| 1353 |
+
"grad_norm": 24.396120071411133,
|
| 1354 |
+
"learning_rate": 6.577650838986077e-06,
|
| 1355 |
+
"loss": 0.7917,
|
| 1356 |
+
"step": 9400
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 3.373795073188147,
|
| 1360 |
+
"grad_norm": 21.905221939086914,
|
| 1361 |
+
"learning_rate": 6.506247768654053e-06,
|
| 1362 |
+
"loss": 0.8816,
|
| 1363 |
+
"step": 9450
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 3.391645840771153,
|
| 1367 |
+
"grad_norm": 17.236255645751953,
|
| 1368 |
+
"learning_rate": 6.434844698322028e-06,
|
| 1369 |
+
"loss": 0.8701,
|
| 1370 |
+
"step": 9500
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 3.4094966083541594,
|
| 1374 |
+
"grad_norm": 17.530790328979492,
|
| 1375 |
+
"learning_rate": 6.363441627990005e-06,
|
| 1376 |
+
"loss": 0.8008,
|
| 1377 |
+
"step": 9550
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 3.427347375937165,
|
| 1381 |
+
"grad_norm": 14.909001350402832,
|
| 1382 |
+
"learning_rate": 6.29203855765798e-06,
|
| 1383 |
+
"loss": 0.8193,
|
| 1384 |
+
"step": 9600
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 3.4451981435201713,
|
| 1388 |
+
"grad_norm": 14.366634368896484,
|
| 1389 |
+
"learning_rate": 6.220635487325956e-06,
|
| 1390 |
+
"loss": 0.8522,
|
| 1391 |
+
"step": 9650
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 3.4630489111031775,
|
| 1395 |
+
"grad_norm": 12.532232284545898,
|
| 1396 |
+
"learning_rate": 6.149232416993931e-06,
|
| 1397 |
+
"loss": 0.927,
|
| 1398 |
+
"step": 9700
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 3.4808996786861837,
|
| 1402 |
+
"grad_norm": 13.383174896240234,
|
| 1403 |
+
"learning_rate": 6.077829346661906e-06,
|
| 1404 |
+
"loss": 0.8455,
|
| 1405 |
+
"step": 9750
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 3.4987504462691894,
|
| 1409 |
+
"grad_norm": 13.50966739654541,
|
| 1410 |
+
"learning_rate": 6.006426276329883e-06,
|
| 1411 |
+
"loss": 0.8659,
|
| 1412 |
+
"step": 9800
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 3.5166012138521956,
|
| 1416 |
+
"grad_norm": 20.433412551879883,
|
| 1417 |
+
"learning_rate": 5.935023205997858e-06,
|
| 1418 |
+
"loss": 0.8791,
|
| 1419 |
+
"step": 9850
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 3.534451981435202,
|
| 1423 |
+
"grad_norm": 15.435810089111328,
|
| 1424 |
+
"learning_rate": 5.863620135665834e-06,
|
| 1425 |
+
"loss": 0.833,
|
| 1426 |
+
"step": 9900
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 3.5523027490182075,
|
| 1430 |
+
"grad_norm": 15.160585403442383,
|
| 1431 |
+
"learning_rate": 5.7922170653338095e-06,
|
| 1432 |
+
"loss": 0.9181,
|
| 1433 |
+
"step": 9950
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 3.5701535166012137,
|
| 1437 |
+
"grad_norm": 24.11473846435547,
|
| 1438 |
+
"learning_rate": 5.720813995001786e-06,
|
| 1439 |
+
"loss": 0.8246,
|
| 1440 |
+
"step": 10000
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 3.58800428418422,
|
| 1444 |
+
"grad_norm": 15.102910995483398,
|
| 1445 |
+
"learning_rate": 5.6494109246697615e-06,
|
| 1446 |
+
"loss": 0.8393,
|
| 1447 |
+
"step": 10050
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 3.605855051767226,
|
| 1451 |
+
"grad_norm": 17.21224021911621,
|
| 1452 |
+
"learning_rate": 5.578007854337737e-06,
|
| 1453 |
+
"loss": 0.8819,
|
| 1454 |
+
"step": 10100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 3.6237058193502323,
|
| 1458 |
+
"grad_norm": 15.384049415588379,
|
| 1459 |
+
"learning_rate": 5.506604784005713e-06,
|
| 1460 |
+
"loss": 0.8691,
|
| 1461 |
+
"step": 10150
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 3.641556586933238,
|
| 1465 |
+
"grad_norm": 16.714080810546875,
|
| 1466 |
+
"learning_rate": 5.4352017136736875e-06,
|
| 1467 |
+
"loss": 0.9156,
|
| 1468 |
+
"step": 10200
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 3.659407354516244,
|
| 1472 |
+
"grad_norm": 12.960479736328125,
|
| 1473 |
+
"learning_rate": 5.363798643341665e-06,
|
| 1474 |
+
"loss": 0.8664,
|
| 1475 |
+
"step": 10250
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 3.6772581220992504,
|
| 1479 |
+
"grad_norm": 18.82040786743164,
|
| 1480 |
+
"learning_rate": 5.2923955730096395e-06,
|
| 1481 |
+
"loss": 0.8805,
|
| 1482 |
+
"step": 10300
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 3.695108889682256,
|
| 1486 |
+
"grad_norm": 16.152734756469727,
|
| 1487 |
+
"learning_rate": 5.220992502677615e-06,
|
| 1488 |
+
"loss": 0.8384,
|
| 1489 |
+
"step": 10350
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 3.7129596572652623,
|
| 1493 |
+
"grad_norm": 15.666060447692871,
|
| 1494 |
+
"learning_rate": 5.149589432345591e-06,
|
| 1495 |
+
"loss": 0.8797,
|
| 1496 |
+
"step": 10400
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 3.7308104248482685,
|
| 1500 |
+
"grad_norm": 15.003888130187988,
|
| 1501 |
+
"learning_rate": 5.078186362013567e-06,
|
| 1502 |
+
"loss": 0.83,
|
| 1503 |
+
"step": 10450
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 3.7486611924312747,
|
| 1507 |
+
"grad_norm": 16.65971565246582,
|
| 1508 |
+
"learning_rate": 5.006783291681543e-06,
|
| 1509 |
+
"loss": 0.8561,
|
| 1510 |
+
"step": 10500
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 3.7665119600142805,
|
| 1514 |
+
"grad_norm": 17.515960693359375,
|
| 1515 |
+
"learning_rate": 4.935380221349518e-06,
|
| 1516 |
+
"loss": 0.8792,
|
| 1517 |
+
"step": 10550
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 3.7843627275972866,
|
| 1521 |
+
"grad_norm": 15.527019500732422,
|
| 1522 |
+
"learning_rate": 4.863977151017494e-06,
|
| 1523 |
+
"loss": 0.8483,
|
| 1524 |
+
"step": 10600
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 3.802213495180293,
|
| 1528 |
+
"grad_norm": 19.536556243896484,
|
| 1529 |
+
"learning_rate": 4.7925740806854695e-06,
|
| 1530 |
+
"loss": 0.8399,
|
| 1531 |
+
"step": 10650
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 3.8200642627632986,
|
| 1535 |
+
"grad_norm": 11.903009414672852,
|
| 1536 |
+
"learning_rate": 4.721171010353446e-06,
|
| 1537 |
+
"loss": 0.8307,
|
| 1538 |
+
"step": 10700
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 3.8379150303463048,
|
| 1542 |
+
"grad_norm": 19.37788200378418,
|
| 1543 |
+
"learning_rate": 4.6497679400214216e-06,
|
| 1544 |
+
"loss": 0.9131,
|
| 1545 |
+
"step": 10750
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 3.855765797929311,
|
| 1549 |
+
"grad_norm": 15.229080200195312,
|
| 1550 |
+
"learning_rate": 4.578364869689397e-06,
|
| 1551 |
+
"loss": 0.8582,
|
| 1552 |
+
"step": 10800
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 3.873616565512317,
|
| 1556 |
+
"grad_norm": 10.229304313659668,
|
| 1557 |
+
"learning_rate": 4.506961799357373e-06,
|
| 1558 |
+
"loss": 0.8102,
|
| 1559 |
+
"step": 10850
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 3.8914673330953233,
|
| 1563 |
+
"grad_norm": 12.341926574707031,
|
| 1564 |
+
"learning_rate": 4.435558729025348e-06,
|
| 1565 |
+
"loss": 0.846,
|
| 1566 |
+
"step": 10900
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 3.909318100678329,
|
| 1570 |
+
"grad_norm": 11.277711868286133,
|
| 1571 |
+
"learning_rate": 4.364155658693324e-06,
|
| 1572 |
+
"loss": 0.8603,
|
| 1573 |
+
"step": 10950
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 3.9271688682613353,
|
| 1577 |
+
"grad_norm": 13.99327278137207,
|
| 1578 |
+
"learning_rate": 4.2927525883612995e-06,
|
| 1579 |
+
"loss": 0.9206,
|
| 1580 |
+
"step": 11000
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 3.9450196358443415,
|
| 1584 |
+
"grad_norm": 25.38481330871582,
|
| 1585 |
+
"learning_rate": 4.221349518029276e-06,
|
| 1586 |
+
"loss": 0.8831,
|
| 1587 |
+
"step": 11050
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 3.962870403427347,
|
| 1591 |
+
"grad_norm": 23.300521850585938,
|
| 1592 |
+
"learning_rate": 4.149946447697252e-06,
|
| 1593 |
+
"loss": 0.8816,
|
| 1594 |
+
"step": 11100
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 3.9807211710103534,
|
| 1598 |
+
"grad_norm": 26.937780380249023,
|
| 1599 |
+
"learning_rate": 4.078543377365227e-06,
|
| 1600 |
+
"loss": 0.8765,
|
| 1601 |
+
"step": 11150
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 3.9985719385933596,
|
| 1605 |
+
"grad_norm": 18.585250854492188,
|
| 1606 |
+
"learning_rate": 4.007140307033203e-06,
|
| 1607 |
+
"loss": 0.7968,
|
| 1608 |
+
"step": 11200
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 4.0,
|
| 1612 |
+
"eval_accuracy": 0.6026957065071856,
|
| 1613 |
+
"eval_loss": 1.3348853588104248,
|
| 1614 |
+
"eval_macro_f1": 0.5141377387367236,
|
| 1615 |
+
"eval_runtime": 5.2222,
|
| 1616 |
+
"eval_samples_per_second": 2145.276,
|
| 1617 |
+
"eval_steps_per_second": 33.702,
|
| 1618 |
+
"step": 11204
|
| 1619 |
+
}
|
| 1620 |
+
],
|
| 1621 |
+
"logging_steps": 50,
|
| 1622 |
+
"max_steps": 14005,
|
| 1623 |
+
"num_input_tokens_seen": 0,
|
| 1624 |
+
"num_train_epochs": 5,
|
| 1625 |
+
"save_steps": 500,
|
| 1626 |
+
"stateful_callbacks": {
|
| 1627 |
+
"EarlyStoppingCallback": {
|
| 1628 |
+
"args": {
|
| 1629 |
+
"early_stopping_patience": 2,
|
| 1630 |
+
"early_stopping_threshold": 0.0
|
| 1631 |
+
},
|
| 1632 |
+
"attributes": {
|
| 1633 |
+
"early_stopping_patience_counter": 0
|
| 1634 |
+
}
|
| 1635 |
+
},
|
| 1636 |
+
"TrainerControl": {
|
| 1637 |
+
"args": {
|
| 1638 |
+
"should_epoch_stop": false,
|
| 1639 |
+
"should_evaluate": false,
|
| 1640 |
+
"should_log": false,
|
| 1641 |
+
"should_save": true,
|
| 1642 |
+
"should_training_stop": false
|
| 1643 |
+
},
|
| 1644 |
+
"attributes": {}
|
| 1645 |
+
}
|
| 1646 |
+
},
|
| 1647 |
+
"total_flos": 5008398947915616.0,
|
| 1648 |
+
"train_batch_size": 32,
|
| 1649 |
+
"trial_name": null,
|
| 1650 |
+
"trial_params": null
|
| 1651 |
+
}
|
checkpoint-11204/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:15055f48195ad08339f81e25f1ac7a0da4c4b8b3553f36d6a79688713db2f8bb
|
| 3 |
+
size 5777
|
checkpoint-11204/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-14005/config.json
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "Aleppo",
|
| 14 |
+
"1": "Alexandria",
|
| 15 |
+
"2": "Algiers",
|
| 16 |
+
"3": "Amman",
|
| 17 |
+
"4": "Aswan",
|
| 18 |
+
"5": "Baghdad",
|
| 19 |
+
"6": "Basra",
|
| 20 |
+
"7": "Beirut",
|
| 21 |
+
"8": "Benghazi",
|
| 22 |
+
"9": "Cairo",
|
| 23 |
+
"10": "Damascus",
|
| 24 |
+
"11": "Doha",
|
| 25 |
+
"12": "Fes",
|
| 26 |
+
"13": "Jeddah",
|
| 27 |
+
"14": "Jerusalem",
|
| 28 |
+
"15": "Khartoum",
|
| 29 |
+
"16": "MSA",
|
| 30 |
+
"17": "Mosul",
|
| 31 |
+
"18": "Muscat",
|
| 32 |
+
"19": "Rabat",
|
| 33 |
+
"20": "Riyadh",
|
| 34 |
+
"21": "Salt",
|
| 35 |
+
"22": "Sanaa",
|
| 36 |
+
"23": "Sfax",
|
| 37 |
+
"24": "Tripoli",
|
| 38 |
+
"25": "Tunis"
|
| 39 |
+
},
|
| 40 |
+
"initializer_range": 0.02,
|
| 41 |
+
"intermediate_size": 3072,
|
| 42 |
+
"label2id": {
|
| 43 |
+
"Aleppo": 0,
|
| 44 |
+
"Alexandria": 1,
|
| 45 |
+
"Algiers": 2,
|
| 46 |
+
"Amman": 3,
|
| 47 |
+
"Aswan": 4,
|
| 48 |
+
"Baghdad": 5,
|
| 49 |
+
"Basra": 6,
|
| 50 |
+
"Beirut": 7,
|
| 51 |
+
"Benghazi": 8,
|
| 52 |
+
"Cairo": 9,
|
| 53 |
+
"Damascus": 10,
|
| 54 |
+
"Doha": 11,
|
| 55 |
+
"Fes": 12,
|
| 56 |
+
"Jeddah": 13,
|
| 57 |
+
"Jerusalem": 14,
|
| 58 |
+
"Khartoum": 15,
|
| 59 |
+
"MSA": 16,
|
| 60 |
+
"Mosul": 17,
|
| 61 |
+
"Muscat": 18,
|
| 62 |
+
"Rabat": 19,
|
| 63 |
+
"Riyadh": 20,
|
| 64 |
+
"Salt": 21,
|
| 65 |
+
"Sanaa": 22,
|
| 66 |
+
"Sfax": 23,
|
| 67 |
+
"Tripoli": 24,
|
| 68 |
+
"Tunis": 25
|
| 69 |
+
},
|
| 70 |
+
"layer_norm_eps": 1e-12,
|
| 71 |
+
"max_position_embeddings": 512,
|
| 72 |
+
"model_type": "bert",
|
| 73 |
+
"num_attention_heads": 12,
|
| 74 |
+
"num_hidden_layers": 12,
|
| 75 |
+
"pad_token_id": 0,
|
| 76 |
+
"position_embedding_type": "absolute",
|
| 77 |
+
"transformers_version": "4.56.2",
|
| 78 |
+
"type_vocab_size": 2,
|
| 79 |
+
"use_cache": true,
|
| 80 |
+
"vocab_size": 30000
|
| 81 |
+
}
|
checkpoint-14005/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e81f9b3ea982c64a859dd00899740789032f2821f6f53fb2d17e3406df66988a
|
| 3 |
+
size 436428880
|
checkpoint-14005/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:48abdf418b119aa7628d89e96ddd610f039f3c7164d4c3748617f988b7beae09
|
| 3 |
+
size 872982027
|
checkpoint-14005/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7211b6bb597b94fbdba035de46beb22b74587dfd510c3a290198996ccf0c809f
|
| 3 |
+
size 14645
|
checkpoint-14005/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:267e16888f66324d8b66a2c89592034b478981272ecc4e21817ae822afa1ffe5
|
| 3 |
+
size 1383
|
checkpoint-14005/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a4b237eb689e34d8a95da580eac8e210da306e5f931ccf39419a98c8200d951d
|
| 3 |
+
size 1465
|
checkpoint-14005/special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
checkpoint-14005/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-14005/tokenizer_config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"full_tokenizer_file": null,
|
| 50 |
+
"mask_token": "[MASK]",
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"never_split": null,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"sep_token": "[SEP]",
|
| 55 |
+
"strip_accents": null,
|
| 56 |
+
"tokenize_chinese_chars": true,
|
| 57 |
+
"tokenizer_class": "BertTokenizer",
|
| 58 |
+
"unk_token": "[UNK]"
|
| 59 |
+
}
|
checkpoint-14005/trainer_state.json
ADDED
|
@@ -0,0 +1,2053 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 14005,
|
| 3 |
+
"best_metric": 0.5237476879491455,
|
| 4 |
+
"best_model_checkpoint": "camelbert_madar_all/checkpoint-14005",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 14005,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.01785076758300607,
|
| 14 |
+
"grad_norm": 12.478485107421875,
|
| 15 |
+
"learning_rate": 1.9930024991074616e-05,
|
| 16 |
+
"loss": 3.29,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.03570153516601214,
|
| 21 |
+
"grad_norm": 8.7147798538208,
|
| 22 |
+
"learning_rate": 1.9858621920742592e-05,
|
| 23 |
+
"loss": 3.0927,
|
| 24 |
+
"step": 100
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.053552302749018205,
|
| 28 |
+
"grad_norm": 9.888656616210938,
|
| 29 |
+
"learning_rate": 1.978721885041057e-05,
|
| 30 |
+
"loss": 2.8678,
|
| 31 |
+
"step": 150
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.07140307033202428,
|
| 35 |
+
"grad_norm": 11.028563499450684,
|
| 36 |
+
"learning_rate": 1.9715815780078543e-05,
|
| 37 |
+
"loss": 2.6757,
|
| 38 |
+
"step": 200
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.08925383791503035,
|
| 42 |
+
"grad_norm": 16.397783279418945,
|
| 43 |
+
"learning_rate": 1.9644412709746522e-05,
|
| 44 |
+
"loss": 2.593,
|
| 45 |
+
"step": 250
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.10710460549803641,
|
| 49 |
+
"grad_norm": 14.620532989501953,
|
| 50 |
+
"learning_rate": 1.9573009639414498e-05,
|
| 51 |
+
"loss": 2.5221,
|
| 52 |
+
"step": 300
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.12495537308104249,
|
| 56 |
+
"grad_norm": 18.156570434570312,
|
| 57 |
+
"learning_rate": 1.9501606569082473e-05,
|
| 58 |
+
"loss": 2.4798,
|
| 59 |
+
"step": 350
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.14280614066404856,
|
| 63 |
+
"grad_norm": 16.055646896362305,
|
| 64 |
+
"learning_rate": 1.943020349875045e-05,
|
| 65 |
+
"loss": 2.4239,
|
| 66 |
+
"step": 400
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.16065690824705461,
|
| 70 |
+
"grad_norm": 16.83954620361328,
|
| 71 |
+
"learning_rate": 1.9358800428418424e-05,
|
| 72 |
+
"loss": 2.3324,
|
| 73 |
+
"step": 450
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.1785076758300607,
|
| 77 |
+
"grad_norm": 11.909624099731445,
|
| 78 |
+
"learning_rate": 1.92873973580864e-05,
|
| 79 |
+
"loss": 2.3261,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.19635844341306677,
|
| 84 |
+
"grad_norm": 11.35456371307373,
|
| 85 |
+
"learning_rate": 1.9215994287754376e-05,
|
| 86 |
+
"loss": 2.1992,
|
| 87 |
+
"step": 550
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.21420921099607282,
|
| 91 |
+
"grad_norm": 13.982519149780273,
|
| 92 |
+
"learning_rate": 1.914459121742235e-05,
|
| 93 |
+
"loss": 2.1815,
|
| 94 |
+
"step": 600
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.2320599785790789,
|
| 98 |
+
"grad_norm": 13.58781623840332,
|
| 99 |
+
"learning_rate": 1.9073188147090327e-05,
|
| 100 |
+
"loss": 2.1739,
|
| 101 |
+
"step": 650
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.24991074616208497,
|
| 105 |
+
"grad_norm": 12.896429061889648,
|
| 106 |
+
"learning_rate": 1.9001785076758302e-05,
|
| 107 |
+
"loss": 2.1063,
|
| 108 |
+
"step": 700
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.26776151374509105,
|
| 112 |
+
"grad_norm": 12.02055549621582,
|
| 113 |
+
"learning_rate": 1.8930382006426278e-05,
|
| 114 |
+
"loss": 2.1043,
|
| 115 |
+
"step": 750
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.28561228132809713,
|
| 119 |
+
"grad_norm": 13.458044052124023,
|
| 120 |
+
"learning_rate": 1.8858978936094254e-05,
|
| 121 |
+
"loss": 2.1073,
|
| 122 |
+
"step": 800
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.30346304891110315,
|
| 126 |
+
"grad_norm": 19.288684844970703,
|
| 127 |
+
"learning_rate": 1.878757586576223e-05,
|
| 128 |
+
"loss": 2.0457,
|
| 129 |
+
"step": 850
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.32131381649410923,
|
| 133 |
+
"grad_norm": 12.411849021911621,
|
| 134 |
+
"learning_rate": 1.8716172795430205e-05,
|
| 135 |
+
"loss": 1.9664,
|
| 136 |
+
"step": 900
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.3391645840771153,
|
| 140 |
+
"grad_norm": 13.870584487915039,
|
| 141 |
+
"learning_rate": 1.864476972509818e-05,
|
| 142 |
+
"loss": 1.9998,
|
| 143 |
+
"step": 950
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.3570153516601214,
|
| 147 |
+
"grad_norm": 9.050592422485352,
|
| 148 |
+
"learning_rate": 1.8573366654766156e-05,
|
| 149 |
+
"loss": 2.0137,
|
| 150 |
+
"step": 1000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.37486611924312746,
|
| 154 |
+
"grad_norm": 10.100810050964355,
|
| 155 |
+
"learning_rate": 1.850196358443413e-05,
|
| 156 |
+
"loss": 1.9402,
|
| 157 |
+
"step": 1050
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.39271688682613354,
|
| 161 |
+
"grad_norm": 15.232287406921387,
|
| 162 |
+
"learning_rate": 1.8430560514102107e-05,
|
| 163 |
+
"loss": 1.8844,
|
| 164 |
+
"step": 1100
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.4105676544091396,
|
| 168 |
+
"grad_norm": 18.118881225585938,
|
| 169 |
+
"learning_rate": 1.8359157443770083e-05,
|
| 170 |
+
"loss": 1.9186,
|
| 171 |
+
"step": 1150
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.42841842199214564,
|
| 175 |
+
"grad_norm": 19.386072158813477,
|
| 176 |
+
"learning_rate": 1.8287754373438062e-05,
|
| 177 |
+
"loss": 1.8201,
|
| 178 |
+
"step": 1200
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.4462691895751517,
|
| 182 |
+
"grad_norm": 15.369712829589844,
|
| 183 |
+
"learning_rate": 1.8216351303106034e-05,
|
| 184 |
+
"loss": 1.8765,
|
| 185 |
+
"step": 1250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.4641199571581578,
|
| 189 |
+
"grad_norm": 16.337072372436523,
|
| 190 |
+
"learning_rate": 1.814494823277401e-05,
|
| 191 |
+
"loss": 1.7933,
|
| 192 |
+
"step": 1300
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.48197072474116387,
|
| 196 |
+
"grad_norm": 12.64062786102295,
|
| 197 |
+
"learning_rate": 1.8073545162441985e-05,
|
| 198 |
+
"loss": 1.8877,
|
| 199 |
+
"step": 1350
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.49982149232416995,
|
| 203 |
+
"grad_norm": 9.410002708435059,
|
| 204 |
+
"learning_rate": 1.800214209210996e-05,
|
| 205 |
+
"loss": 1.8393,
|
| 206 |
+
"step": 1400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.517672259907176,
|
| 210 |
+
"grad_norm": 14.48257827758789,
|
| 211 |
+
"learning_rate": 1.793073902177794e-05,
|
| 212 |
+
"loss": 1.7939,
|
| 213 |
+
"step": 1450
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.5355230274901821,
|
| 217 |
+
"grad_norm": 14.2874116897583,
|
| 218 |
+
"learning_rate": 1.7859335951445912e-05,
|
| 219 |
+
"loss": 1.8296,
|
| 220 |
+
"step": 1500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.5533737950731882,
|
| 224 |
+
"grad_norm": 13.599530220031738,
|
| 225 |
+
"learning_rate": 1.778793288111389e-05,
|
| 226 |
+
"loss": 1.8267,
|
| 227 |
+
"step": 1550
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.5712245626561943,
|
| 231 |
+
"grad_norm": 15.281228065490723,
|
| 232 |
+
"learning_rate": 1.7716529810781867e-05,
|
| 233 |
+
"loss": 1.8124,
|
| 234 |
+
"step": 1600
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.5890753302392003,
|
| 238 |
+
"grad_norm": 16.20423126220703,
|
| 239 |
+
"learning_rate": 1.7645126740449842e-05,
|
| 240 |
+
"loss": 1.7258,
|
| 241 |
+
"step": 1650
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.6069260978222063,
|
| 245 |
+
"grad_norm": 17.073945999145508,
|
| 246 |
+
"learning_rate": 1.7573723670117818e-05,
|
| 247 |
+
"loss": 1.7866,
|
| 248 |
+
"step": 1700
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.6247768654052124,
|
| 252 |
+
"grad_norm": 12.85055160522461,
|
| 253 |
+
"learning_rate": 1.750232059978579e-05,
|
| 254 |
+
"loss": 1.7171,
|
| 255 |
+
"step": 1750
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.6426276329882185,
|
| 259 |
+
"grad_norm": 26.679119110107422,
|
| 260 |
+
"learning_rate": 1.743091752945377e-05,
|
| 261 |
+
"loss": 1.7145,
|
| 262 |
+
"step": 1800
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.6604784005712245,
|
| 266 |
+
"grad_norm": 20.80165672302246,
|
| 267 |
+
"learning_rate": 1.7359514459121744e-05,
|
| 268 |
+
"loss": 1.7033,
|
| 269 |
+
"step": 1850
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.6783291681542306,
|
| 273 |
+
"grad_norm": 12.617931365966797,
|
| 274 |
+
"learning_rate": 1.728811138878972e-05,
|
| 275 |
+
"loss": 1.6911,
|
| 276 |
+
"step": 1900
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.6961799357372367,
|
| 280 |
+
"grad_norm": 12.220250129699707,
|
| 281 |
+
"learning_rate": 1.7216708318457696e-05,
|
| 282 |
+
"loss": 1.7364,
|
| 283 |
+
"step": 1950
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.7140307033202428,
|
| 287 |
+
"grad_norm": 12.190744400024414,
|
| 288 |
+
"learning_rate": 1.714530524812567e-05,
|
| 289 |
+
"loss": 1.6801,
|
| 290 |
+
"step": 2000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.7318814709032488,
|
| 294 |
+
"grad_norm": 14.35610580444336,
|
| 295 |
+
"learning_rate": 1.7073902177793647e-05,
|
| 296 |
+
"loss": 1.6561,
|
| 297 |
+
"step": 2050
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.7497322384862549,
|
| 301 |
+
"grad_norm": 18.017112731933594,
|
| 302 |
+
"learning_rate": 1.7002499107461622e-05,
|
| 303 |
+
"loss": 1.6541,
|
| 304 |
+
"step": 2100
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.767583006069261,
|
| 308 |
+
"grad_norm": 17.52274513244629,
|
| 309 |
+
"learning_rate": 1.6931096037129598e-05,
|
| 310 |
+
"loss": 1.6398,
|
| 311 |
+
"step": 2150
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.7854337736522671,
|
| 315 |
+
"grad_norm": 10.50312328338623,
|
| 316 |
+
"learning_rate": 1.6859692966797574e-05,
|
| 317 |
+
"loss": 1.5567,
|
| 318 |
+
"step": 2200
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.8032845412352732,
|
| 322 |
+
"grad_norm": 13.701729774475098,
|
| 323 |
+
"learning_rate": 1.678828989646555e-05,
|
| 324 |
+
"loss": 1.5603,
|
| 325 |
+
"step": 2250
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.8211353088182792,
|
| 329 |
+
"grad_norm": 11.937254905700684,
|
| 330 |
+
"learning_rate": 1.6716886826133525e-05,
|
| 331 |
+
"loss": 1.6716,
|
| 332 |
+
"step": 2300
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.8389860764012853,
|
| 336 |
+
"grad_norm": 17.40386199951172,
|
| 337 |
+
"learning_rate": 1.66454837558015e-05,
|
| 338 |
+
"loss": 1.6773,
|
| 339 |
+
"step": 2350
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.8568368439842913,
|
| 343 |
+
"grad_norm": 15.239500045776367,
|
| 344 |
+
"learning_rate": 1.6574080685469476e-05,
|
| 345 |
+
"loss": 1.596,
|
| 346 |
+
"step": 2400
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.8746876115672974,
|
| 350 |
+
"grad_norm": 14.679793357849121,
|
| 351 |
+
"learning_rate": 1.650267761513745e-05,
|
| 352 |
+
"loss": 1.5855,
|
| 353 |
+
"step": 2450
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.8925383791503034,
|
| 357 |
+
"grad_norm": 12.250946998596191,
|
| 358 |
+
"learning_rate": 1.643127454480543e-05,
|
| 359 |
+
"loss": 1.6029,
|
| 360 |
+
"step": 2500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.9103891467333095,
|
| 364 |
+
"grad_norm": 14.957499504089355,
|
| 365 |
+
"learning_rate": 1.6359871474473403e-05,
|
| 366 |
+
"loss": 1.6466,
|
| 367 |
+
"step": 2550
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.9282399143163156,
|
| 371 |
+
"grad_norm": 21.005565643310547,
|
| 372 |
+
"learning_rate": 1.628846840414138e-05,
|
| 373 |
+
"loss": 1.5891,
|
| 374 |
+
"step": 2600
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.9460906818993217,
|
| 378 |
+
"grad_norm": 13.731078147888184,
|
| 379 |
+
"learning_rate": 1.6217065333809354e-05,
|
| 380 |
+
"loss": 1.5669,
|
| 381 |
+
"step": 2650
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.9639414494823277,
|
| 385 |
+
"grad_norm": 14.127930641174316,
|
| 386 |
+
"learning_rate": 1.614566226347733e-05,
|
| 387 |
+
"loss": 1.5489,
|
| 388 |
+
"step": 2700
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.9817922170653338,
|
| 392 |
+
"grad_norm": 12.521357536315918,
|
| 393 |
+
"learning_rate": 1.607425919314531e-05,
|
| 394 |
+
"loss": 1.5861,
|
| 395 |
+
"step": 2750
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.9996429846483399,
|
| 399 |
+
"grad_norm": 12.26593017578125,
|
| 400 |
+
"learning_rate": 1.600285612281328e-05,
|
| 401 |
+
"loss": 1.4877,
|
| 402 |
+
"step": 2800
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.0,
|
| 406 |
+
"eval_accuracy": 0.47514058734267606,
|
| 407 |
+
"eval_loss": 1.4969408512115479,
|
| 408 |
+
"eval_macro_f1": 0.4136955039973382,
|
| 409 |
+
"eval_runtime": 5.1653,
|
| 410 |
+
"eval_samples_per_second": 2168.895,
|
| 411 |
+
"eval_steps_per_second": 34.074,
|
| 412 |
+
"step": 2801
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.0174937522313459,
|
| 416 |
+
"grad_norm": 15.36621379852295,
|
| 417 |
+
"learning_rate": 1.593145305248126e-05,
|
| 418 |
+
"loss": 1.4006,
|
| 419 |
+
"step": 2850
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.035344519814352,
|
| 423 |
+
"grad_norm": 12.056408882141113,
|
| 424 |
+
"learning_rate": 1.5860049982149232e-05,
|
| 425 |
+
"loss": 1.3621,
|
| 426 |
+
"step": 2900
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.053195287397358,
|
| 430 |
+
"grad_norm": 10.554118156433105,
|
| 431 |
+
"learning_rate": 1.5788646911817208e-05,
|
| 432 |
+
"loss": 1.3599,
|
| 433 |
+
"step": 2950
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 1.0710460549803642,
|
| 437 |
+
"grad_norm": 13.228151321411133,
|
| 438 |
+
"learning_rate": 1.5717243841485187e-05,
|
| 439 |
+
"loss": 1.4433,
|
| 440 |
+
"step": 3000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 1.0888968225633702,
|
| 444 |
+
"grad_norm": 19.01808738708496,
|
| 445 |
+
"learning_rate": 1.564584077115316e-05,
|
| 446 |
+
"loss": 1.3759,
|
| 447 |
+
"step": 3050
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 1.1067475901463764,
|
| 451 |
+
"grad_norm": 12.861035346984863,
|
| 452 |
+
"learning_rate": 1.5574437700821138e-05,
|
| 453 |
+
"loss": 1.4442,
|
| 454 |
+
"step": 3100
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 1.1245983577293823,
|
| 458 |
+
"grad_norm": 9.24783706665039,
|
| 459 |
+
"learning_rate": 1.5503034630489113e-05,
|
| 460 |
+
"loss": 1.3682,
|
| 461 |
+
"step": 3150
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.1424491253123885,
|
| 465 |
+
"grad_norm": 15.15107250213623,
|
| 466 |
+
"learning_rate": 1.543163156015709e-05,
|
| 467 |
+
"loss": 1.3637,
|
| 468 |
+
"step": 3200
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 1.1602998928953945,
|
| 472 |
+
"grad_norm": 12.833767890930176,
|
| 473 |
+
"learning_rate": 1.5360228489825065e-05,
|
| 474 |
+
"loss": 1.3372,
|
| 475 |
+
"step": 3250
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 1.1781506604784004,
|
| 479 |
+
"grad_norm": 12.872199058532715,
|
| 480 |
+
"learning_rate": 1.528882541949304e-05,
|
| 481 |
+
"loss": 1.4134,
|
| 482 |
+
"step": 3300
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.1960014280614066,
|
| 486 |
+
"grad_norm": 12.539731979370117,
|
| 487 |
+
"learning_rate": 1.5217422349161016e-05,
|
| 488 |
+
"loss": 1.3883,
|
| 489 |
+
"step": 3350
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.2138521956444128,
|
| 493 |
+
"grad_norm": 14.508119583129883,
|
| 494 |
+
"learning_rate": 1.5146019278828991e-05,
|
| 495 |
+
"loss": 1.3427,
|
| 496 |
+
"step": 3400
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.2317029632274188,
|
| 500 |
+
"grad_norm": 18.304662704467773,
|
| 501 |
+
"learning_rate": 1.5074616208496965e-05,
|
| 502 |
+
"loss": 1.3978,
|
| 503 |
+
"step": 3450
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.2495537308104248,
|
| 507 |
+
"grad_norm": 15.10910701751709,
|
| 508 |
+
"learning_rate": 1.5003213138164943e-05,
|
| 509 |
+
"loss": 1.389,
|
| 510 |
+
"step": 3500
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.267404498393431,
|
| 514 |
+
"grad_norm": 12.175374984741211,
|
| 515 |
+
"learning_rate": 1.4931810067832916e-05,
|
| 516 |
+
"loss": 1.2985,
|
| 517 |
+
"step": 3550
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 1.285255265976437,
|
| 521 |
+
"grad_norm": 11.14751148223877,
|
| 522 |
+
"learning_rate": 1.4860406997500894e-05,
|
| 523 |
+
"loss": 1.384,
|
| 524 |
+
"step": 3600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.303106033559443,
|
| 528 |
+
"grad_norm": 15.333986282348633,
|
| 529 |
+
"learning_rate": 1.478900392716887e-05,
|
| 530 |
+
"loss": 1.2794,
|
| 531 |
+
"step": 3650
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.320956801142449,
|
| 535 |
+
"grad_norm": 15.89001178741455,
|
| 536 |
+
"learning_rate": 1.4717600856836845e-05,
|
| 537 |
+
"loss": 1.2597,
|
| 538 |
+
"step": 3700
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 1.3388075687254553,
|
| 542 |
+
"grad_norm": 9.596504211425781,
|
| 543 |
+
"learning_rate": 1.464619778650482e-05,
|
| 544 |
+
"loss": 1.358,
|
| 545 |
+
"step": 3750
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.3566583363084612,
|
| 549 |
+
"grad_norm": 12.051373481750488,
|
| 550 |
+
"learning_rate": 1.4574794716172798e-05,
|
| 551 |
+
"loss": 1.3429,
|
| 552 |
+
"step": 3800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.3745091038914674,
|
| 556 |
+
"grad_norm": 9.407623291015625,
|
| 557 |
+
"learning_rate": 1.4503391645840772e-05,
|
| 558 |
+
"loss": 1.3347,
|
| 559 |
+
"step": 3850
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.3923598714744734,
|
| 563 |
+
"grad_norm": 13.37045669555664,
|
| 564 |
+
"learning_rate": 1.4431988575508749e-05,
|
| 565 |
+
"loss": 1.3214,
|
| 566 |
+
"step": 3900
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.4102106390574796,
|
| 570 |
+
"grad_norm": 14.173185348510742,
|
| 571 |
+
"learning_rate": 1.4360585505176723e-05,
|
| 572 |
+
"loss": 1.3509,
|
| 573 |
+
"step": 3950
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.4280614066404855,
|
| 577 |
+
"grad_norm": 20.702144622802734,
|
| 578 |
+
"learning_rate": 1.42891824348447e-05,
|
| 579 |
+
"loss": 1.3057,
|
| 580 |
+
"step": 4000
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.4459121742234915,
|
| 584 |
+
"grad_norm": 11.873905181884766,
|
| 585 |
+
"learning_rate": 1.4217779364512676e-05,
|
| 586 |
+
"loss": 1.333,
|
| 587 |
+
"step": 4050
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.4637629418064977,
|
| 591 |
+
"grad_norm": 12.91700553894043,
|
| 592 |
+
"learning_rate": 1.414637629418065e-05,
|
| 593 |
+
"loss": 1.2759,
|
| 594 |
+
"step": 4100
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.4816137093895039,
|
| 598 |
+
"grad_norm": 16.117185592651367,
|
| 599 |
+
"learning_rate": 1.4074973223848627e-05,
|
| 600 |
+
"loss": 1.2325,
|
| 601 |
+
"step": 4150
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.4994644769725098,
|
| 605 |
+
"grad_norm": 12.514154434204102,
|
| 606 |
+
"learning_rate": 1.40035701535166e-05,
|
| 607 |
+
"loss": 1.3223,
|
| 608 |
+
"step": 4200
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.5173152445555158,
|
| 612 |
+
"grad_norm": 19.87337303161621,
|
| 613 |
+
"learning_rate": 1.3932167083184578e-05,
|
| 614 |
+
"loss": 1.3427,
|
| 615 |
+
"step": 4250
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.535166012138522,
|
| 619 |
+
"grad_norm": 13.149452209472656,
|
| 620 |
+
"learning_rate": 1.3860764012852554e-05,
|
| 621 |
+
"loss": 1.3193,
|
| 622 |
+
"step": 4300
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 1.5530167797215282,
|
| 626 |
+
"grad_norm": 16.208499908447266,
|
| 627 |
+
"learning_rate": 1.378936094252053e-05,
|
| 628 |
+
"loss": 1.3719,
|
| 629 |
+
"step": 4350
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 1.570867547304534,
|
| 633 |
+
"grad_norm": 15.929703712463379,
|
| 634 |
+
"learning_rate": 1.3717957872188505e-05,
|
| 635 |
+
"loss": 1.369,
|
| 636 |
+
"step": 4400
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 1.5887183148875401,
|
| 640 |
+
"grad_norm": 12.900412559509277,
|
| 641 |
+
"learning_rate": 1.3646554801856482e-05,
|
| 642 |
+
"loss": 1.3369,
|
| 643 |
+
"step": 4450
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.6065690824705463,
|
| 647 |
+
"grad_norm": 13.354098320007324,
|
| 648 |
+
"learning_rate": 1.3575151731524456e-05,
|
| 649 |
+
"loss": 1.2681,
|
| 650 |
+
"step": 4500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.6244198500535523,
|
| 654 |
+
"grad_norm": 17.960058212280273,
|
| 655 |
+
"learning_rate": 1.3503748661192433e-05,
|
| 656 |
+
"loss": 1.2589,
|
| 657 |
+
"step": 4550
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.6422706176365582,
|
| 661 |
+
"grad_norm": 17.655719757080078,
|
| 662 |
+
"learning_rate": 1.3432345590860407e-05,
|
| 663 |
+
"loss": 1.3318,
|
| 664 |
+
"step": 4600
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 1.6601213852195644,
|
| 668 |
+
"grad_norm": 19.73931121826172,
|
| 669 |
+
"learning_rate": 1.3360942520528385e-05,
|
| 670 |
+
"loss": 1.2895,
|
| 671 |
+
"step": 4650
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 1.6779721528025706,
|
| 675 |
+
"grad_norm": 14.323142051696777,
|
| 676 |
+
"learning_rate": 1.328953945019636e-05,
|
| 677 |
+
"loss": 1.2691,
|
| 678 |
+
"step": 4700
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 1.6958229203855766,
|
| 682 |
+
"grad_norm": 8.690425872802734,
|
| 683 |
+
"learning_rate": 1.3218136379864334e-05,
|
| 684 |
+
"loss": 1.2606,
|
| 685 |
+
"step": 4750
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.7136736879685825,
|
| 689 |
+
"grad_norm": 10.671348571777344,
|
| 690 |
+
"learning_rate": 1.3146733309532311e-05,
|
| 691 |
+
"loss": 1.3214,
|
| 692 |
+
"step": 4800
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.7315244555515887,
|
| 696 |
+
"grad_norm": 8.343477249145508,
|
| 697 |
+
"learning_rate": 1.3075330239200285e-05,
|
| 698 |
+
"loss": 1.2785,
|
| 699 |
+
"step": 4850
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.749375223134595,
|
| 703 |
+
"grad_norm": 22.884681701660156,
|
| 704 |
+
"learning_rate": 1.3003927168868263e-05,
|
| 705 |
+
"loss": 1.2526,
|
| 706 |
+
"step": 4900
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 1.767225990717601,
|
| 710 |
+
"grad_norm": 17.500192642211914,
|
| 711 |
+
"learning_rate": 1.2932524098536238e-05,
|
| 712 |
+
"loss": 1.2738,
|
| 713 |
+
"step": 4950
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 1.7850767583006069,
|
| 717 |
+
"grad_norm": 17.966121673583984,
|
| 718 |
+
"learning_rate": 1.2861121028204214e-05,
|
| 719 |
+
"loss": 1.2983,
|
| 720 |
+
"step": 5000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 1.802927525883613,
|
| 724 |
+
"grad_norm": 22.048490524291992,
|
| 725 |
+
"learning_rate": 1.278971795787219e-05,
|
| 726 |
+
"loss": 1.3437,
|
| 727 |
+
"step": 5050
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.820778293466619,
|
| 731 |
+
"grad_norm": 16.436044692993164,
|
| 732 |
+
"learning_rate": 1.2718314887540167e-05,
|
| 733 |
+
"loss": 1.2604,
|
| 734 |
+
"step": 5100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.838629061049625,
|
| 738 |
+
"grad_norm": 12.748188972473145,
|
| 739 |
+
"learning_rate": 1.264691181720814e-05,
|
| 740 |
+
"loss": 1.1955,
|
| 741 |
+
"step": 5150
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 1.8564798286326312,
|
| 745 |
+
"grad_norm": 15.320619583129883,
|
| 746 |
+
"learning_rate": 1.2575508746876118e-05,
|
| 747 |
+
"loss": 1.2902,
|
| 748 |
+
"step": 5200
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.8743305962156374,
|
| 752 |
+
"grad_norm": 17.871013641357422,
|
| 753 |
+
"learning_rate": 1.2504105676544092e-05,
|
| 754 |
+
"loss": 1.3103,
|
| 755 |
+
"step": 5250
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.8921813637986433,
|
| 759 |
+
"grad_norm": 16.096118927001953,
|
| 760 |
+
"learning_rate": 1.2432702606212067e-05,
|
| 761 |
+
"loss": 1.3319,
|
| 762 |
+
"step": 5300
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.9100321313816493,
|
| 766 |
+
"grad_norm": 15.915545463562012,
|
| 767 |
+
"learning_rate": 1.2361299535880045e-05,
|
| 768 |
+
"loss": 1.3093,
|
| 769 |
+
"step": 5350
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.9278828989646555,
|
| 773 |
+
"grad_norm": 12.922966003417969,
|
| 774 |
+
"learning_rate": 1.2289896465548019e-05,
|
| 775 |
+
"loss": 1.2533,
|
| 776 |
+
"step": 5400
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.9457336665476617,
|
| 780 |
+
"grad_norm": 20.724853515625,
|
| 781 |
+
"learning_rate": 1.2218493395215996e-05,
|
| 782 |
+
"loss": 1.2531,
|
| 783 |
+
"step": 5450
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 1.9635844341306676,
|
| 787 |
+
"grad_norm": 13.744089126586914,
|
| 788 |
+
"learning_rate": 1.214709032488397e-05,
|
| 789 |
+
"loss": 1.2476,
|
| 790 |
+
"step": 5500
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.9814352017136736,
|
| 794 |
+
"grad_norm": 11.663731575012207,
|
| 795 |
+
"learning_rate": 1.2075687254551947e-05,
|
| 796 |
+
"loss": 1.2768,
|
| 797 |
+
"step": 5550
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 1.9992859692966798,
|
| 801 |
+
"grad_norm": 16.515056610107422,
|
| 802 |
+
"learning_rate": 1.2004284184219923e-05,
|
| 803 |
+
"loss": 1.283,
|
| 804 |
+
"step": 5600
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 2.0,
|
| 808 |
+
"eval_accuracy": 0.5697581005087923,
|
| 809 |
+
"eval_loss": 1.3204034566879272,
|
| 810 |
+
"eval_macro_f1": 0.48569754853380576,
|
| 811 |
+
"eval_runtime": 5.4967,
|
| 812 |
+
"eval_samples_per_second": 2038.126,
|
| 813 |
+
"eval_steps_per_second": 32.019,
|
| 814 |
+
"step": 5602
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 2.017136736879686,
|
| 818 |
+
"grad_norm": 10.937089920043945,
|
| 819 |
+
"learning_rate": 1.1932881113887898e-05,
|
| 820 |
+
"loss": 1.0781,
|
| 821 |
+
"step": 5650
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 2.0349875044626917,
|
| 825 |
+
"grad_norm": 17.1431827545166,
|
| 826 |
+
"learning_rate": 1.1861478043555874e-05,
|
| 827 |
+
"loss": 1.034,
|
| 828 |
+
"step": 5700
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 2.052838272045698,
|
| 832 |
+
"grad_norm": 12.69800853729248,
|
| 833 |
+
"learning_rate": 1.1790074973223851e-05,
|
| 834 |
+
"loss": 1.0675,
|
| 835 |
+
"step": 5750
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 2.070689039628704,
|
| 839 |
+
"grad_norm": 19.09876251220703,
|
| 840 |
+
"learning_rate": 1.1718671902891825e-05,
|
| 841 |
+
"loss": 1.0352,
|
| 842 |
+
"step": 5800
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 2.0885398072117103,
|
| 846 |
+
"grad_norm": 16.88023567199707,
|
| 847 |
+
"learning_rate": 1.1647268832559802e-05,
|
| 848 |
+
"loss": 0.9982,
|
| 849 |
+
"step": 5850
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.106390574794716,
|
| 853 |
+
"grad_norm": 9.332680702209473,
|
| 854 |
+
"learning_rate": 1.1575865762227776e-05,
|
| 855 |
+
"loss": 1.058,
|
| 856 |
+
"step": 5900
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 2.124241342377722,
|
| 860 |
+
"grad_norm": 15.302008628845215,
|
| 861 |
+
"learning_rate": 1.1504462691895752e-05,
|
| 862 |
+
"loss": 1.0775,
|
| 863 |
+
"step": 5950
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 2.1420921099607284,
|
| 867 |
+
"grad_norm": 13.956599235534668,
|
| 868 |
+
"learning_rate": 1.1433059621563729e-05,
|
| 869 |
+
"loss": 1.0847,
|
| 870 |
+
"step": 6000
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 2.1599428775437346,
|
| 874 |
+
"grad_norm": 13.09730052947998,
|
| 875 |
+
"learning_rate": 1.1361656551231703e-05,
|
| 876 |
+
"loss": 1.0621,
|
| 877 |
+
"step": 6050
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 2.1777936451267403,
|
| 881 |
+
"grad_norm": 18.93950080871582,
|
| 882 |
+
"learning_rate": 1.129025348089968e-05,
|
| 883 |
+
"loss": 1.0375,
|
| 884 |
+
"step": 6100
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.1956444127097465,
|
| 888 |
+
"grad_norm": 16.003028869628906,
|
| 889 |
+
"learning_rate": 1.1218850410567654e-05,
|
| 890 |
+
"loss": 1.0133,
|
| 891 |
+
"step": 6150
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.2134951802927527,
|
| 895 |
+
"grad_norm": 11.415363311767578,
|
| 896 |
+
"learning_rate": 1.1147447340235631e-05,
|
| 897 |
+
"loss": 1.0313,
|
| 898 |
+
"step": 6200
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 2.2313459478757585,
|
| 902 |
+
"grad_norm": 15.419634819030762,
|
| 903 |
+
"learning_rate": 1.1076044269903607e-05,
|
| 904 |
+
"loss": 1.0583,
|
| 905 |
+
"step": 6250
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.2491967154587647,
|
| 909 |
+
"grad_norm": 16.330265045166016,
|
| 910 |
+
"learning_rate": 1.1004641199571583e-05,
|
| 911 |
+
"loss": 0.9851,
|
| 912 |
+
"step": 6300
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 2.267047483041771,
|
| 916 |
+
"grad_norm": 12.965764999389648,
|
| 917 |
+
"learning_rate": 1.0933238129239558e-05,
|
| 918 |
+
"loss": 1.0314,
|
| 919 |
+
"step": 6350
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 2.284898250624777,
|
| 923 |
+
"grad_norm": 13.175607681274414,
|
| 924 |
+
"learning_rate": 1.0861835058907535e-05,
|
| 925 |
+
"loss": 1.0458,
|
| 926 |
+
"step": 6400
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.3027490182077828,
|
| 930 |
+
"grad_norm": 17.50844383239746,
|
| 931 |
+
"learning_rate": 1.079043198857551e-05,
|
| 932 |
+
"loss": 1.0705,
|
| 933 |
+
"step": 6450
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.320599785790789,
|
| 937 |
+
"grad_norm": 11.109663009643555,
|
| 938 |
+
"learning_rate": 1.0719028918243487e-05,
|
| 939 |
+
"loss": 1.1261,
|
| 940 |
+
"step": 6500
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 2.338450553373795,
|
| 944 |
+
"grad_norm": 22.36432456970215,
|
| 945 |
+
"learning_rate": 1.064762584791146e-05,
|
| 946 |
+
"loss": 1.0871,
|
| 947 |
+
"step": 6550
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 2.356301320956801,
|
| 951 |
+
"grad_norm": 12.304158210754395,
|
| 952 |
+
"learning_rate": 1.0576222777579436e-05,
|
| 953 |
+
"loss": 1.0667,
|
| 954 |
+
"step": 6600
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 2.374152088539807,
|
| 958 |
+
"grad_norm": 17.145753860473633,
|
| 959 |
+
"learning_rate": 1.0504819707247413e-05,
|
| 960 |
+
"loss": 1.0315,
|
| 961 |
+
"step": 6650
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 2.3920028561228133,
|
| 965 |
+
"grad_norm": 10.1165189743042,
|
| 966 |
+
"learning_rate": 1.0433416636915387e-05,
|
| 967 |
+
"loss": 1.0643,
|
| 968 |
+
"step": 6700
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.4098536237058195,
|
| 972 |
+
"grad_norm": 11.35006332397461,
|
| 973 |
+
"learning_rate": 1.0362013566583365e-05,
|
| 974 |
+
"loss": 1.0621,
|
| 975 |
+
"step": 6750
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 2.4277043912888256,
|
| 979 |
+
"grad_norm": 13.226170539855957,
|
| 980 |
+
"learning_rate": 1.0290610496251339e-05,
|
| 981 |
+
"loss": 1.0086,
|
| 982 |
+
"step": 6800
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 2.4455551588718314,
|
| 986 |
+
"grad_norm": 13.296606063842773,
|
| 987 |
+
"learning_rate": 1.0219207425919316e-05,
|
| 988 |
+
"loss": 1.0329,
|
| 989 |
+
"step": 6850
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 2.4634059264548376,
|
| 993 |
+
"grad_norm": 11.565988540649414,
|
| 994 |
+
"learning_rate": 1.0147804355587291e-05,
|
| 995 |
+
"loss": 1.0779,
|
| 996 |
+
"step": 6900
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 2.4812566940378438,
|
| 1000 |
+
"grad_norm": 21.814311981201172,
|
| 1001 |
+
"learning_rate": 1.0076401285255265e-05,
|
| 1002 |
+
"loss": 1.0901,
|
| 1003 |
+
"step": 6950
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 2.4991074616208495,
|
| 1007 |
+
"grad_norm": 20.675832748413086,
|
| 1008 |
+
"learning_rate": 1.0004998214923243e-05,
|
| 1009 |
+
"loss": 1.0447,
|
| 1010 |
+
"step": 7000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.5169582292038557,
|
| 1014 |
+
"grad_norm": 18.146474838256836,
|
| 1015 |
+
"learning_rate": 9.933595144591218e-06,
|
| 1016 |
+
"loss": 1.1045,
|
| 1017 |
+
"step": 7050
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.534808996786862,
|
| 1021 |
+
"grad_norm": 14.326371192932129,
|
| 1022 |
+
"learning_rate": 9.862192074259194e-06,
|
| 1023 |
+
"loss": 1.0594,
|
| 1024 |
+
"step": 7100
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 2.552659764369868,
|
| 1028 |
+
"grad_norm": 14.145441055297852,
|
| 1029 |
+
"learning_rate": 9.79078900392717e-06,
|
| 1030 |
+
"loss": 1.0402,
|
| 1031 |
+
"step": 7150
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 2.570510531952874,
|
| 1035 |
+
"grad_norm": 17.572402954101562,
|
| 1036 |
+
"learning_rate": 9.719385933595145e-06,
|
| 1037 |
+
"loss": 1.0297,
|
| 1038 |
+
"step": 7200
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 2.58836129953588,
|
| 1042 |
+
"grad_norm": 16.921300888061523,
|
| 1043 |
+
"learning_rate": 9.64798286326312e-06,
|
| 1044 |
+
"loss": 1.0694,
|
| 1045 |
+
"step": 7250
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 2.606212067118886,
|
| 1049 |
+
"grad_norm": 9.290846824645996,
|
| 1050 |
+
"learning_rate": 9.576579792931096e-06,
|
| 1051 |
+
"loss": 1.0014,
|
| 1052 |
+
"step": 7300
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.624062834701892,
|
| 1056 |
+
"grad_norm": 18.461746215820312,
|
| 1057 |
+
"learning_rate": 9.505176722599072e-06,
|
| 1058 |
+
"loss": 1.0958,
|
| 1059 |
+
"step": 7350
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 2.641913602284898,
|
| 1063 |
+
"grad_norm": 18.638286590576172,
|
| 1064 |
+
"learning_rate": 9.433773652267049e-06,
|
| 1065 |
+
"loss": 1.0575,
|
| 1066 |
+
"step": 7400
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 2.6597643698679043,
|
| 1070 |
+
"grad_norm": 20.554609298706055,
|
| 1071 |
+
"learning_rate": 9.362370581935025e-06,
|
| 1072 |
+
"loss": 1.0293,
|
| 1073 |
+
"step": 7450
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.6776151374509105,
|
| 1077 |
+
"grad_norm": 17.582111358642578,
|
| 1078 |
+
"learning_rate": 9.290967511603e-06,
|
| 1079 |
+
"loss": 1.0127,
|
| 1080 |
+
"step": 7500
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.6954659050339167,
|
| 1084 |
+
"grad_norm": 18.36084747314453,
|
| 1085 |
+
"learning_rate": 9.219564441270976e-06,
|
| 1086 |
+
"loss": 1.0402,
|
| 1087 |
+
"step": 7550
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 2.7133166726169224,
|
| 1091 |
+
"grad_norm": 11.184069633483887,
|
| 1092 |
+
"learning_rate": 9.148161370938951e-06,
|
| 1093 |
+
"loss": 1.0198,
|
| 1094 |
+
"step": 7600
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.7311674401999286,
|
| 1098 |
+
"grad_norm": 19.872299194335938,
|
| 1099 |
+
"learning_rate": 9.076758300606927e-06,
|
| 1100 |
+
"loss": 1.0562,
|
| 1101 |
+
"step": 7650
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 2.749018207782935,
|
| 1105 |
+
"grad_norm": 18.872329711914062,
|
| 1106 |
+
"learning_rate": 9.005355230274903e-06,
|
| 1107 |
+
"loss": 1.0907,
|
| 1108 |
+
"step": 7700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 2.7668689753659406,
|
| 1112 |
+
"grad_norm": 21.884220123291016,
|
| 1113 |
+
"learning_rate": 8.933952159942878e-06,
|
| 1114 |
+
"loss": 1.0713,
|
| 1115 |
+
"step": 7750
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.7847197429489468,
|
| 1119 |
+
"grad_norm": 20.70819854736328,
|
| 1120 |
+
"learning_rate": 8.862549089610854e-06,
|
| 1121 |
+
"loss": 1.001,
|
| 1122 |
+
"step": 7800
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 2.802570510531953,
|
| 1126 |
+
"grad_norm": 33.434959411621094,
|
| 1127 |
+
"learning_rate": 8.79114601927883e-06,
|
| 1128 |
+
"loss": 1.0503,
|
| 1129 |
+
"step": 7850
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 2.820421278114959,
|
| 1133 |
+
"grad_norm": 16.19710922241211,
|
| 1134 |
+
"learning_rate": 8.719742948946805e-06,
|
| 1135 |
+
"loss": 1.0701,
|
| 1136 |
+
"step": 7900
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.838272045697965,
|
| 1140 |
+
"grad_norm": 14.586867332458496,
|
| 1141 |
+
"learning_rate": 8.64833987861478e-06,
|
| 1142 |
+
"loss": 1.0619,
|
| 1143 |
+
"step": 7950
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 2.856122813280971,
|
| 1147 |
+
"grad_norm": 15.705036163330078,
|
| 1148 |
+
"learning_rate": 8.576936808282756e-06,
|
| 1149 |
+
"loss": 1.0708,
|
| 1150 |
+
"step": 8000
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 2.8739735808639773,
|
| 1154 |
+
"grad_norm": 14.276032447814941,
|
| 1155 |
+
"learning_rate": 8.505533737950734e-06,
|
| 1156 |
+
"loss": 1.1002,
|
| 1157 |
+
"step": 8050
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.891824348446983,
|
| 1161 |
+
"grad_norm": 19.014434814453125,
|
| 1162 |
+
"learning_rate": 8.434130667618709e-06,
|
| 1163 |
+
"loss": 0.9907,
|
| 1164 |
+
"step": 8100
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 2.909675116029989,
|
| 1168 |
+
"grad_norm": 20.64139747619629,
|
| 1169 |
+
"learning_rate": 8.362727597286685e-06,
|
| 1170 |
+
"loss": 1.0514,
|
| 1171 |
+
"step": 8150
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 2.9275258836129954,
|
| 1175 |
+
"grad_norm": 19.546451568603516,
|
| 1176 |
+
"learning_rate": 8.291324526954659e-06,
|
| 1177 |
+
"loss": 1.059,
|
| 1178 |
+
"step": 8200
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.9453766511960016,
|
| 1182 |
+
"grad_norm": 18.8275203704834,
|
| 1183 |
+
"learning_rate": 8.219921456622636e-06,
|
| 1184 |
+
"loss": 1.0642,
|
| 1185 |
+
"step": 8250
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 2.9632274187790077,
|
| 1189 |
+
"grad_norm": 14.960760116577148,
|
| 1190 |
+
"learning_rate": 8.148518386290611e-06,
|
| 1191 |
+
"loss": 1.1098,
|
| 1192 |
+
"step": 8300
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 2.9810781863620135,
|
| 1196 |
+
"grad_norm": 11.512022972106934,
|
| 1197 |
+
"learning_rate": 8.077115315958587e-06,
|
| 1198 |
+
"loss": 1.0295,
|
| 1199 |
+
"step": 8350
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.9989289539450197,
|
| 1203 |
+
"grad_norm": 18.68143653869629,
|
| 1204 |
+
"learning_rate": 8.005712245626563e-06,
|
| 1205 |
+
"loss": 1.0402,
|
| 1206 |
+
"step": 8400
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 3.0,
|
| 1210 |
+
"eval_accuracy": 0.5899312684102472,
|
| 1211 |
+
"eval_loss": 1.2934809923171997,
|
| 1212 |
+
"eval_macro_f1": 0.5093257341608811,
|
| 1213 |
+
"eval_runtime": 5.2479,
|
| 1214 |
+
"eval_samples_per_second": 2134.777,
|
| 1215 |
+
"eval_steps_per_second": 33.538,
|
| 1216 |
+
"step": 8403
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 3.016779721528026,
|
| 1220 |
+
"grad_norm": 19.454416275024414,
|
| 1221 |
+
"learning_rate": 7.934309175294538e-06,
|
| 1222 |
+
"loss": 0.8671,
|
| 1223 |
+
"step": 8450
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 3.0346304891110316,
|
| 1227 |
+
"grad_norm": 14.609670639038086,
|
| 1228 |
+
"learning_rate": 7.862906104962514e-06,
|
| 1229 |
+
"loss": 0.8195,
|
| 1230 |
+
"step": 8500
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 3.052481256694038,
|
| 1234 |
+
"grad_norm": 13.611706733703613,
|
| 1235 |
+
"learning_rate": 7.79150303463049e-06,
|
| 1236 |
+
"loss": 0.9039,
|
| 1237 |
+
"step": 8550
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 3.070332024277044,
|
| 1241 |
+
"grad_norm": 17.289657592773438,
|
| 1242 |
+
"learning_rate": 7.720099964298465e-06,
|
| 1243 |
+
"loss": 0.8817,
|
| 1244 |
+
"step": 8600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 3.08818279186005,
|
| 1248 |
+
"grad_norm": 12.592385292053223,
|
| 1249 |
+
"learning_rate": 7.64869689396644e-06,
|
| 1250 |
+
"loss": 0.8339,
|
| 1251 |
+
"step": 8650
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 3.106033559443056,
|
| 1255 |
+
"grad_norm": 17.638628005981445,
|
| 1256 |
+
"learning_rate": 7.577293823634417e-06,
|
| 1257 |
+
"loss": 0.8151,
|
| 1258 |
+
"step": 8700
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 3.123884327026062,
|
| 1262 |
+
"grad_norm": 11.680763244628906,
|
| 1263 |
+
"learning_rate": 7.505890753302393e-06,
|
| 1264 |
+
"loss": 0.8438,
|
| 1265 |
+
"step": 8750
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 3.1417350946090683,
|
| 1269 |
+
"grad_norm": 22.238727569580078,
|
| 1270 |
+
"learning_rate": 7.434487682970368e-06,
|
| 1271 |
+
"loss": 0.8263,
|
| 1272 |
+
"step": 8800
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 3.159585862192074,
|
| 1276 |
+
"grad_norm": 13.042167663574219,
|
| 1277 |
+
"learning_rate": 7.363084612638344e-06,
|
| 1278 |
+
"loss": 0.8538,
|
| 1279 |
+
"step": 8850
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 3.1774366297750802,
|
| 1283 |
+
"grad_norm": 11.398730278015137,
|
| 1284 |
+
"learning_rate": 7.29168154230632e-06,
|
| 1285 |
+
"loss": 0.8465,
|
| 1286 |
+
"step": 8900
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 3.1952873973580864,
|
| 1290 |
+
"grad_norm": 12.321819305419922,
|
| 1291 |
+
"learning_rate": 7.220278471974296e-06,
|
| 1292 |
+
"loss": 0.8297,
|
| 1293 |
+
"step": 8950
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 3.2131381649410926,
|
| 1297 |
+
"grad_norm": 10.085368156433105,
|
| 1298 |
+
"learning_rate": 7.1488754016422715e-06,
|
| 1299 |
+
"loss": 0.8655,
|
| 1300 |
+
"step": 9000
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 3.2309889325240984,
|
| 1304 |
+
"grad_norm": 7.707581043243408,
|
| 1305 |
+
"learning_rate": 7.077472331310247e-06,
|
| 1306 |
+
"loss": 0.7967,
|
| 1307 |
+
"step": 9050
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 3.2488397001071045,
|
| 1311 |
+
"grad_norm": 13.6712646484375,
|
| 1312 |
+
"learning_rate": 7.006069260978222e-06,
|
| 1313 |
+
"loss": 0.8858,
|
| 1314 |
+
"step": 9100
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 3.2666904676901107,
|
| 1318 |
+
"grad_norm": 24.826183319091797,
|
| 1319 |
+
"learning_rate": 6.934666190646199e-06,
|
| 1320 |
+
"loss": 0.8799,
|
| 1321 |
+
"step": 9150
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 3.284541235273117,
|
| 1325 |
+
"grad_norm": 16.819063186645508,
|
| 1326 |
+
"learning_rate": 6.863263120314174e-06,
|
| 1327 |
+
"loss": 0.889,
|
| 1328 |
+
"step": 9200
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 3.3023920028561227,
|
| 1332 |
+
"grad_norm": 16.014638900756836,
|
| 1333 |
+
"learning_rate": 6.7918600499821495e-06,
|
| 1334 |
+
"loss": 0.8766,
|
| 1335 |
+
"step": 9250
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 3.320242770439129,
|
| 1339 |
+
"grad_norm": 16.722623825073242,
|
| 1340 |
+
"learning_rate": 6.720456979650125e-06,
|
| 1341 |
+
"loss": 0.8391,
|
| 1342 |
+
"step": 9300
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 3.338093538022135,
|
| 1346 |
+
"grad_norm": 15.532989501953125,
|
| 1347 |
+
"learning_rate": 6.6490539093181015e-06,
|
| 1348 |
+
"loss": 0.906,
|
| 1349 |
+
"step": 9350
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 3.3559443056051412,
|
| 1353 |
+
"grad_norm": 24.396120071411133,
|
| 1354 |
+
"learning_rate": 6.577650838986077e-06,
|
| 1355 |
+
"loss": 0.7917,
|
| 1356 |
+
"step": 9400
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 3.373795073188147,
|
| 1360 |
+
"grad_norm": 21.905221939086914,
|
| 1361 |
+
"learning_rate": 6.506247768654053e-06,
|
| 1362 |
+
"loss": 0.8816,
|
| 1363 |
+
"step": 9450
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 3.391645840771153,
|
| 1367 |
+
"grad_norm": 17.236255645751953,
|
| 1368 |
+
"learning_rate": 6.434844698322028e-06,
|
| 1369 |
+
"loss": 0.8701,
|
| 1370 |
+
"step": 9500
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 3.4094966083541594,
|
| 1374 |
+
"grad_norm": 17.530790328979492,
|
| 1375 |
+
"learning_rate": 6.363441627990005e-06,
|
| 1376 |
+
"loss": 0.8008,
|
| 1377 |
+
"step": 9550
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 3.427347375937165,
|
| 1381 |
+
"grad_norm": 14.909001350402832,
|
| 1382 |
+
"learning_rate": 6.29203855765798e-06,
|
| 1383 |
+
"loss": 0.8193,
|
| 1384 |
+
"step": 9600
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 3.4451981435201713,
|
| 1388 |
+
"grad_norm": 14.366634368896484,
|
| 1389 |
+
"learning_rate": 6.220635487325956e-06,
|
| 1390 |
+
"loss": 0.8522,
|
| 1391 |
+
"step": 9650
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 3.4630489111031775,
|
| 1395 |
+
"grad_norm": 12.532232284545898,
|
| 1396 |
+
"learning_rate": 6.149232416993931e-06,
|
| 1397 |
+
"loss": 0.927,
|
| 1398 |
+
"step": 9700
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 3.4808996786861837,
|
| 1402 |
+
"grad_norm": 13.383174896240234,
|
| 1403 |
+
"learning_rate": 6.077829346661906e-06,
|
| 1404 |
+
"loss": 0.8455,
|
| 1405 |
+
"step": 9750
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 3.4987504462691894,
|
| 1409 |
+
"grad_norm": 13.50966739654541,
|
| 1410 |
+
"learning_rate": 6.006426276329883e-06,
|
| 1411 |
+
"loss": 0.8659,
|
| 1412 |
+
"step": 9800
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 3.5166012138521956,
|
| 1416 |
+
"grad_norm": 20.433412551879883,
|
| 1417 |
+
"learning_rate": 5.935023205997858e-06,
|
| 1418 |
+
"loss": 0.8791,
|
| 1419 |
+
"step": 9850
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 3.534451981435202,
|
| 1423 |
+
"grad_norm": 15.435810089111328,
|
| 1424 |
+
"learning_rate": 5.863620135665834e-06,
|
| 1425 |
+
"loss": 0.833,
|
| 1426 |
+
"step": 9900
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 3.5523027490182075,
|
| 1430 |
+
"grad_norm": 15.160585403442383,
|
| 1431 |
+
"learning_rate": 5.7922170653338095e-06,
|
| 1432 |
+
"loss": 0.9181,
|
| 1433 |
+
"step": 9950
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 3.5701535166012137,
|
| 1437 |
+
"grad_norm": 24.11473846435547,
|
| 1438 |
+
"learning_rate": 5.720813995001786e-06,
|
| 1439 |
+
"loss": 0.8246,
|
| 1440 |
+
"step": 10000
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 3.58800428418422,
|
| 1444 |
+
"grad_norm": 15.102910995483398,
|
| 1445 |
+
"learning_rate": 5.6494109246697615e-06,
|
| 1446 |
+
"loss": 0.8393,
|
| 1447 |
+
"step": 10050
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 3.605855051767226,
|
| 1451 |
+
"grad_norm": 17.21224021911621,
|
| 1452 |
+
"learning_rate": 5.578007854337737e-06,
|
| 1453 |
+
"loss": 0.8819,
|
| 1454 |
+
"step": 10100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 3.6237058193502323,
|
| 1458 |
+
"grad_norm": 15.384049415588379,
|
| 1459 |
+
"learning_rate": 5.506604784005713e-06,
|
| 1460 |
+
"loss": 0.8691,
|
| 1461 |
+
"step": 10150
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 3.641556586933238,
|
| 1465 |
+
"grad_norm": 16.714080810546875,
|
| 1466 |
+
"learning_rate": 5.4352017136736875e-06,
|
| 1467 |
+
"loss": 0.9156,
|
| 1468 |
+
"step": 10200
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 3.659407354516244,
|
| 1472 |
+
"grad_norm": 12.960479736328125,
|
| 1473 |
+
"learning_rate": 5.363798643341665e-06,
|
| 1474 |
+
"loss": 0.8664,
|
| 1475 |
+
"step": 10250
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 3.6772581220992504,
|
| 1479 |
+
"grad_norm": 18.82040786743164,
|
| 1480 |
+
"learning_rate": 5.2923955730096395e-06,
|
| 1481 |
+
"loss": 0.8805,
|
| 1482 |
+
"step": 10300
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 3.695108889682256,
|
| 1486 |
+
"grad_norm": 16.152734756469727,
|
| 1487 |
+
"learning_rate": 5.220992502677615e-06,
|
| 1488 |
+
"loss": 0.8384,
|
| 1489 |
+
"step": 10350
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 3.7129596572652623,
|
| 1493 |
+
"grad_norm": 15.666060447692871,
|
| 1494 |
+
"learning_rate": 5.149589432345591e-06,
|
| 1495 |
+
"loss": 0.8797,
|
| 1496 |
+
"step": 10400
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 3.7308104248482685,
|
| 1500 |
+
"grad_norm": 15.003888130187988,
|
| 1501 |
+
"learning_rate": 5.078186362013567e-06,
|
| 1502 |
+
"loss": 0.83,
|
| 1503 |
+
"step": 10450
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 3.7486611924312747,
|
| 1507 |
+
"grad_norm": 16.65971565246582,
|
| 1508 |
+
"learning_rate": 5.006783291681543e-06,
|
| 1509 |
+
"loss": 0.8561,
|
| 1510 |
+
"step": 10500
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 3.7665119600142805,
|
| 1514 |
+
"grad_norm": 17.515960693359375,
|
| 1515 |
+
"learning_rate": 4.935380221349518e-06,
|
| 1516 |
+
"loss": 0.8792,
|
| 1517 |
+
"step": 10550
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 3.7843627275972866,
|
| 1521 |
+
"grad_norm": 15.527019500732422,
|
| 1522 |
+
"learning_rate": 4.863977151017494e-06,
|
| 1523 |
+
"loss": 0.8483,
|
| 1524 |
+
"step": 10600
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 3.802213495180293,
|
| 1528 |
+
"grad_norm": 19.536556243896484,
|
| 1529 |
+
"learning_rate": 4.7925740806854695e-06,
|
| 1530 |
+
"loss": 0.8399,
|
| 1531 |
+
"step": 10650
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 3.8200642627632986,
|
| 1535 |
+
"grad_norm": 11.903009414672852,
|
| 1536 |
+
"learning_rate": 4.721171010353446e-06,
|
| 1537 |
+
"loss": 0.8307,
|
| 1538 |
+
"step": 10700
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 3.8379150303463048,
|
| 1542 |
+
"grad_norm": 19.37788200378418,
|
| 1543 |
+
"learning_rate": 4.6497679400214216e-06,
|
| 1544 |
+
"loss": 0.9131,
|
| 1545 |
+
"step": 10750
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 3.855765797929311,
|
| 1549 |
+
"grad_norm": 15.229080200195312,
|
| 1550 |
+
"learning_rate": 4.578364869689397e-06,
|
| 1551 |
+
"loss": 0.8582,
|
| 1552 |
+
"step": 10800
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 3.873616565512317,
|
| 1556 |
+
"grad_norm": 10.229304313659668,
|
| 1557 |
+
"learning_rate": 4.506961799357373e-06,
|
| 1558 |
+
"loss": 0.8102,
|
| 1559 |
+
"step": 10850
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 3.8914673330953233,
|
| 1563 |
+
"grad_norm": 12.341926574707031,
|
| 1564 |
+
"learning_rate": 4.435558729025348e-06,
|
| 1565 |
+
"loss": 0.846,
|
| 1566 |
+
"step": 10900
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 3.909318100678329,
|
| 1570 |
+
"grad_norm": 11.277711868286133,
|
| 1571 |
+
"learning_rate": 4.364155658693324e-06,
|
| 1572 |
+
"loss": 0.8603,
|
| 1573 |
+
"step": 10950
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 3.9271688682613353,
|
| 1577 |
+
"grad_norm": 13.99327278137207,
|
| 1578 |
+
"learning_rate": 4.2927525883612995e-06,
|
| 1579 |
+
"loss": 0.9206,
|
| 1580 |
+
"step": 11000
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 3.9450196358443415,
|
| 1584 |
+
"grad_norm": 25.38481330871582,
|
| 1585 |
+
"learning_rate": 4.221349518029276e-06,
|
| 1586 |
+
"loss": 0.8831,
|
| 1587 |
+
"step": 11050
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 3.962870403427347,
|
| 1591 |
+
"grad_norm": 23.300521850585938,
|
| 1592 |
+
"learning_rate": 4.149946447697252e-06,
|
| 1593 |
+
"loss": 0.8816,
|
| 1594 |
+
"step": 11100
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 3.9807211710103534,
|
| 1598 |
+
"grad_norm": 26.937780380249023,
|
| 1599 |
+
"learning_rate": 4.078543377365227e-06,
|
| 1600 |
+
"loss": 0.8765,
|
| 1601 |
+
"step": 11150
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 3.9985719385933596,
|
| 1605 |
+
"grad_norm": 18.585250854492188,
|
| 1606 |
+
"learning_rate": 4.007140307033203e-06,
|
| 1607 |
+
"loss": 0.7968,
|
| 1608 |
+
"step": 11200
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 4.0,
|
| 1612 |
+
"eval_accuracy": 0.6026957065071856,
|
| 1613 |
+
"eval_loss": 1.3348853588104248,
|
| 1614 |
+
"eval_macro_f1": 0.5141377387367236,
|
| 1615 |
+
"eval_runtime": 5.2222,
|
| 1616 |
+
"eval_samples_per_second": 2145.276,
|
| 1617 |
+
"eval_steps_per_second": 33.702,
|
| 1618 |
+
"step": 11204
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 4.016422706176366,
|
| 1622 |
+
"grad_norm": 21.511226654052734,
|
| 1623 |
+
"learning_rate": 3.935737236701178e-06,
|
| 1624 |
+
"loss": 0.6859,
|
| 1625 |
+
"step": 11250
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 4.034273473759372,
|
| 1629 |
+
"grad_norm": 14.637051582336426,
|
| 1630 |
+
"learning_rate": 3.864334166369154e-06,
|
| 1631 |
+
"loss": 0.6551,
|
| 1632 |
+
"step": 11300
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 4.052124241342378,
|
| 1636 |
+
"grad_norm": 8.740469932556152,
|
| 1637 |
+
"learning_rate": 3.79293109603713e-06,
|
| 1638 |
+
"loss": 0.7377,
|
| 1639 |
+
"step": 11350
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 4.069975008925383,
|
| 1643 |
+
"grad_norm": 20.587265014648438,
|
| 1644 |
+
"learning_rate": 3.7215280257051056e-06,
|
| 1645 |
+
"loss": 0.6943,
|
| 1646 |
+
"step": 11400
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 4.08782577650839,
|
| 1650 |
+
"grad_norm": 25.303258895874023,
|
| 1651 |
+
"learning_rate": 3.650124955373081e-06,
|
| 1652 |
+
"loss": 0.6796,
|
| 1653 |
+
"step": 11450
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 4.105676544091396,
|
| 1657 |
+
"grad_norm": 31.2967586517334,
|
| 1658 |
+
"learning_rate": 3.578721885041057e-06,
|
| 1659 |
+
"loss": 0.7483,
|
| 1660 |
+
"step": 11500
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 4.123527311674402,
|
| 1664 |
+
"grad_norm": 12.077474594116211,
|
| 1665 |
+
"learning_rate": 3.5073188147090328e-06,
|
| 1666 |
+
"loss": 0.7165,
|
| 1667 |
+
"step": 11550
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 4.141378079257408,
|
| 1671 |
+
"grad_norm": 13.995460510253906,
|
| 1672 |
+
"learning_rate": 3.435915744377009e-06,
|
| 1673 |
+
"loss": 0.6947,
|
| 1674 |
+
"step": 11600
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 4.159228846840414,
|
| 1678 |
+
"grad_norm": 14.387024879455566,
|
| 1679 |
+
"learning_rate": 3.364512674044984e-06,
|
| 1680 |
+
"loss": 0.6874,
|
| 1681 |
+
"step": 11650
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 4.177079614423421,
|
| 1685 |
+
"grad_norm": 14.671749114990234,
|
| 1686 |
+
"learning_rate": 3.29310960371296e-06,
|
| 1687 |
+
"loss": 0.7658,
|
| 1688 |
+
"step": 11700
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 4.194930382006426,
|
| 1692 |
+
"grad_norm": 19.197551727294922,
|
| 1693 |
+
"learning_rate": 3.2217065333809356e-06,
|
| 1694 |
+
"loss": 0.6566,
|
| 1695 |
+
"step": 11750
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 4.212781149589432,
|
| 1699 |
+
"grad_norm": 24.187671661376953,
|
| 1700 |
+
"learning_rate": 3.1503034630489116e-06,
|
| 1701 |
+
"loss": 0.7452,
|
| 1702 |
+
"step": 11800
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 4.230631917172438,
|
| 1706 |
+
"grad_norm": 14.38919448852539,
|
| 1707 |
+
"learning_rate": 3.078900392716887e-06,
|
| 1708 |
+
"loss": 0.7136,
|
| 1709 |
+
"step": 11850
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 4.248482684755444,
|
| 1713 |
+
"grad_norm": 18.66862678527832,
|
| 1714 |
+
"learning_rate": 3.0074973223848632e-06,
|
| 1715 |
+
"loss": 0.7537,
|
| 1716 |
+
"step": 11900
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 4.266333452338451,
|
| 1720 |
+
"grad_norm": 13.695405960083008,
|
| 1721 |
+
"learning_rate": 2.9360942520528384e-06,
|
| 1722 |
+
"loss": 0.7418,
|
| 1723 |
+
"step": 11950
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 4.284184219921457,
|
| 1727 |
+
"grad_norm": 19.877029418945312,
|
| 1728 |
+
"learning_rate": 2.8646911817208144e-06,
|
| 1729 |
+
"loss": 0.7052,
|
| 1730 |
+
"step": 12000
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 4.302034987504463,
|
| 1734 |
+
"grad_norm": 15.04555892944336,
|
| 1735 |
+
"learning_rate": 2.79328811138879e-06,
|
| 1736 |
+
"loss": 0.7,
|
| 1737 |
+
"step": 12050
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 4.319885755087469,
|
| 1741 |
+
"grad_norm": 11.697660446166992,
|
| 1742 |
+
"learning_rate": 2.7218850410567656e-06,
|
| 1743 |
+
"loss": 0.7305,
|
| 1744 |
+
"step": 12100
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 4.3377365226704745,
|
| 1748 |
+
"grad_norm": 10.877930641174316,
|
| 1749 |
+
"learning_rate": 2.6504819707247416e-06,
|
| 1750 |
+
"loss": 0.7064,
|
| 1751 |
+
"step": 12150
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 4.355587290253481,
|
| 1755 |
+
"grad_norm": 19.296728134155273,
|
| 1756 |
+
"learning_rate": 2.5790789003927168e-06,
|
| 1757 |
+
"loss": 0.74,
|
| 1758 |
+
"step": 12200
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 4.373438057836487,
|
| 1762 |
+
"grad_norm": 18.58037567138672,
|
| 1763 |
+
"learning_rate": 2.507675830060693e-06,
|
| 1764 |
+
"loss": 0.7146,
|
| 1765 |
+
"step": 12250
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 4.391288825419493,
|
| 1769 |
+
"grad_norm": 12.399540901184082,
|
| 1770 |
+
"learning_rate": 2.436272759728669e-06,
|
| 1771 |
+
"loss": 0.7369,
|
| 1772 |
+
"step": 12300
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 4.409139593002499,
|
| 1776 |
+
"grad_norm": 21.00368309020996,
|
| 1777 |
+
"learning_rate": 2.364869689396644e-06,
|
| 1778 |
+
"loss": 0.7005,
|
| 1779 |
+
"step": 12350
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 4.426990360585505,
|
| 1783 |
+
"grad_norm": 13.065181732177734,
|
| 1784 |
+
"learning_rate": 2.29346661906462e-06,
|
| 1785 |
+
"loss": 0.7207,
|
| 1786 |
+
"step": 12400
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 4.444841128168512,
|
| 1790 |
+
"grad_norm": 15.18995189666748,
|
| 1791 |
+
"learning_rate": 2.2220635487325956e-06,
|
| 1792 |
+
"loss": 0.6926,
|
| 1793 |
+
"step": 12450
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 4.462691895751517,
|
| 1797 |
+
"grad_norm": 13.790769577026367,
|
| 1798 |
+
"learning_rate": 2.150660478400571e-06,
|
| 1799 |
+
"loss": 0.7483,
|
| 1800 |
+
"step": 12500
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 4.480542663334523,
|
| 1804 |
+
"grad_norm": 15.466925621032715,
|
| 1805 |
+
"learning_rate": 2.0792574080685472e-06,
|
| 1806 |
+
"loss": 0.7481,
|
| 1807 |
+
"step": 12550
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 4.498393430917529,
|
| 1811 |
+
"grad_norm": 19.100069046020508,
|
| 1812 |
+
"learning_rate": 2.007854337736523e-06,
|
| 1813 |
+
"loss": 0.7462,
|
| 1814 |
+
"step": 12600
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 4.5162441985005355,
|
| 1818 |
+
"grad_norm": 16.695058822631836,
|
| 1819 |
+
"learning_rate": 1.9364512674044984e-06,
|
| 1820 |
+
"loss": 0.7259,
|
| 1821 |
+
"step": 12650
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 4.534094966083542,
|
| 1825 |
+
"grad_norm": 13.062968254089355,
|
| 1826 |
+
"learning_rate": 1.8650481970724742e-06,
|
| 1827 |
+
"loss": 0.7803,
|
| 1828 |
+
"step": 12700
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 4.551945733666548,
|
| 1832 |
+
"grad_norm": 23.521682739257812,
|
| 1833 |
+
"learning_rate": 1.79364512674045e-06,
|
| 1834 |
+
"loss": 0.7229,
|
| 1835 |
+
"step": 12750
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 4.569796501249554,
|
| 1839 |
+
"grad_norm": 11.35993480682373,
|
| 1840 |
+
"learning_rate": 1.7222420564084258e-06,
|
| 1841 |
+
"loss": 0.6811,
|
| 1842 |
+
"step": 12800
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 4.58764726883256,
|
| 1846 |
+
"grad_norm": 20.912187576293945,
|
| 1847 |
+
"learning_rate": 1.6508389860764014e-06,
|
| 1848 |
+
"loss": 0.7502,
|
| 1849 |
+
"step": 12850
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 4.6054980364155655,
|
| 1853 |
+
"grad_norm": 12.022528648376465,
|
| 1854 |
+
"learning_rate": 1.5794359157443772e-06,
|
| 1855 |
+
"loss": 0.7371,
|
| 1856 |
+
"step": 12900
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 4.623348803998572,
|
| 1860 |
+
"grad_norm": 22.161956787109375,
|
| 1861 |
+
"learning_rate": 1.5080328454123528e-06,
|
| 1862 |
+
"loss": 0.7027,
|
| 1863 |
+
"step": 12950
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 4.641199571581578,
|
| 1867 |
+
"grad_norm": 14.684049606323242,
|
| 1868 |
+
"learning_rate": 1.4366297750803284e-06,
|
| 1869 |
+
"loss": 0.758,
|
| 1870 |
+
"step": 13000
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 4.659050339164584,
|
| 1874 |
+
"grad_norm": 21.676809310913086,
|
| 1875 |
+
"learning_rate": 1.3652267047483042e-06,
|
| 1876 |
+
"loss": 0.7714,
|
| 1877 |
+
"step": 13050
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 4.67690110674759,
|
| 1881 |
+
"grad_norm": 19.062238693237305,
|
| 1882 |
+
"learning_rate": 1.29382363441628e-06,
|
| 1883 |
+
"loss": 0.6905,
|
| 1884 |
+
"step": 13100
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 4.6947518743305965,
|
| 1888 |
+
"grad_norm": 10.306020736694336,
|
| 1889 |
+
"learning_rate": 1.2224205640842556e-06,
|
| 1890 |
+
"loss": 0.7415,
|
| 1891 |
+
"step": 13150
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 4.712602641913602,
|
| 1895 |
+
"grad_norm": 15.594263076782227,
|
| 1896 |
+
"learning_rate": 1.1510174937522314e-06,
|
| 1897 |
+
"loss": 0.6667,
|
| 1898 |
+
"step": 13200
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 4.730453409496608,
|
| 1902 |
+
"grad_norm": 25.428884506225586,
|
| 1903 |
+
"learning_rate": 1.0796144234202072e-06,
|
| 1904 |
+
"loss": 0.7354,
|
| 1905 |
+
"step": 13250
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 4.748304177079614,
|
| 1909 |
+
"grad_norm": 14.635579109191895,
|
| 1910 |
+
"learning_rate": 1.0082113530881828e-06,
|
| 1911 |
+
"loss": 0.6969,
|
| 1912 |
+
"step": 13300
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 4.76615494466262,
|
| 1916 |
+
"grad_norm": 10.086297988891602,
|
| 1917 |
+
"learning_rate": 9.368082827561586e-07,
|
| 1918 |
+
"loss": 0.7221,
|
| 1919 |
+
"step": 13350
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 4.7840057122456265,
|
| 1923 |
+
"grad_norm": 14.498841285705566,
|
| 1924 |
+
"learning_rate": 8.654052124241342e-07,
|
| 1925 |
+
"loss": 0.7424,
|
| 1926 |
+
"step": 13400
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 4.801856479828633,
|
| 1930 |
+
"grad_norm": 20.77524757385254,
|
| 1931 |
+
"learning_rate": 7.940021420921099e-07,
|
| 1932 |
+
"loss": 0.7295,
|
| 1933 |
+
"step": 13450
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 4.819707247411639,
|
| 1937 |
+
"grad_norm": 14.695490837097168,
|
| 1938 |
+
"learning_rate": 7.225990717600857e-07,
|
| 1939 |
+
"loss": 0.6597,
|
| 1940 |
+
"step": 13500
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 4.837558014994645,
|
| 1944 |
+
"grad_norm": 11.918572425842285,
|
| 1945 |
+
"learning_rate": 6.511960014280614e-07,
|
| 1946 |
+
"loss": 0.7785,
|
| 1947 |
+
"step": 13550
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 4.855408782577651,
|
| 1951 |
+
"grad_norm": 13.884369850158691,
|
| 1952 |
+
"learning_rate": 5.797929310960373e-07,
|
| 1953 |
+
"loss": 0.6682,
|
| 1954 |
+
"step": 13600
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 4.873259550160657,
|
| 1958 |
+
"grad_norm": 14.705413818359375,
|
| 1959 |
+
"learning_rate": 5.083898607640128e-07,
|
| 1960 |
+
"loss": 0.6688,
|
| 1961 |
+
"step": 13650
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 4.891110317743663,
|
| 1965 |
+
"grad_norm": 13.439905166625977,
|
| 1966 |
+
"learning_rate": 4.369867904319886e-07,
|
| 1967 |
+
"loss": 0.7471,
|
| 1968 |
+
"step": 13700
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 4.908961085326669,
|
| 1972 |
+
"grad_norm": 26.38125228881836,
|
| 1973 |
+
"learning_rate": 3.6558372009996436e-07,
|
| 1974 |
+
"loss": 0.7365,
|
| 1975 |
+
"step": 13750
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 4.926811852909675,
|
| 1979 |
+
"grad_norm": 18.915802001953125,
|
| 1980 |
+
"learning_rate": 2.9418064976794e-07,
|
| 1981 |
+
"loss": 0.7168,
|
| 1982 |
+
"step": 13800
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 4.944662620492681,
|
| 1986 |
+
"grad_norm": 18.118656158447266,
|
| 1987 |
+
"learning_rate": 2.2277757943591576e-07,
|
| 1988 |
+
"loss": 0.7313,
|
| 1989 |
+
"step": 13850
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 4.9625133880756875,
|
| 1993 |
+
"grad_norm": 16.01120948791504,
|
| 1994 |
+
"learning_rate": 1.5137450910389149e-07,
|
| 1995 |
+
"loss": 0.7383,
|
| 1996 |
+
"step": 13900
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 4.980364155658693,
|
| 2000 |
+
"grad_norm": 14.877579689025879,
|
| 2001 |
+
"learning_rate": 7.99714387718672e-08,
|
| 2002 |
+
"loss": 0.7174,
|
| 2003 |
+
"step": 13950
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 4.998214923241699,
|
| 2007 |
+
"grad_norm": 16.274520874023438,
|
| 2008 |
+
"learning_rate": 8.568368439842914e-09,
|
| 2009 |
+
"loss": 0.7552,
|
| 2010 |
+
"step": 14000
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 5.0,
|
| 2014 |
+
"eval_accuracy": 0.6225118271891458,
|
| 2015 |
+
"eval_loss": 1.3671221733093262,
|
| 2016 |
+
"eval_macro_f1": 0.5237476879491455,
|
| 2017 |
+
"eval_runtime": 5.2302,
|
| 2018 |
+
"eval_samples_per_second": 2141.976,
|
| 2019 |
+
"eval_steps_per_second": 33.651,
|
| 2020 |
+
"step": 14005
|
| 2021 |
+
}
|
| 2022 |
+
],
|
| 2023 |
+
"logging_steps": 50,
|
| 2024 |
+
"max_steps": 14005,
|
| 2025 |
+
"num_input_tokens_seen": 0,
|
| 2026 |
+
"num_train_epochs": 5,
|
| 2027 |
+
"save_steps": 500,
|
| 2028 |
+
"stateful_callbacks": {
|
| 2029 |
+
"EarlyStoppingCallback": {
|
| 2030 |
+
"args": {
|
| 2031 |
+
"early_stopping_patience": 2,
|
| 2032 |
+
"early_stopping_threshold": 0.0
|
| 2033 |
+
},
|
| 2034 |
+
"attributes": {
|
| 2035 |
+
"early_stopping_patience_counter": 0
|
| 2036 |
+
}
|
| 2037 |
+
},
|
| 2038 |
+
"TrainerControl": {
|
| 2039 |
+
"args": {
|
| 2040 |
+
"should_epoch_stop": false,
|
| 2041 |
+
"should_evaluate": false,
|
| 2042 |
+
"should_log": false,
|
| 2043 |
+
"should_save": true,
|
| 2044 |
+
"should_training_stop": true
|
| 2045 |
+
},
|
| 2046 |
+
"attributes": {}
|
| 2047 |
+
}
|
| 2048 |
+
},
|
| 2049 |
+
"total_flos": 6261169968262416.0,
|
| 2050 |
+
"train_batch_size": 32,
|
| 2051 |
+
"trial_name": null,
|
| 2052 |
+
"trial_params": null
|
| 2053 |
+
}
|
checkpoint-14005/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:15055f48195ad08339f81e25f1ac7a0da4c4b8b3553f36d6a79688713db2f8bb
|
| 3 |
+
size 5777
|
checkpoint-14005/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
metrics.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"timestamp": "2025-10-07T19:13:15.820867Z",
|
| 3 |
+
"dev": {
|
| 4 |
+
"eval_loss": 1.3671221733093262,
|
| 5 |
+
"eval_accuracy": 0.6225118271891458,
|
| 6 |
+
"eval_macro_f1": 0.5237476879491455,
|
| 7 |
+
"eval_runtime": 5.2939,
|
| 8 |
+
"eval_samples_per_second": 2116.218,
|
| 9 |
+
"eval_steps_per_second": 33.246,
|
| 10 |
+
"epoch": 5.0
|
| 11 |
+
},
|
| 12 |
+
"test": {
|
| 13 |
+
"eval_loss": 1.4015007019042969,
|
| 14 |
+
"eval_accuracy": 0.61367490850665,
|
| 15 |
+
"eval_macro_f1": 0.515946969897074,
|
| 16 |
+
"eval_runtime": 5.5094,
|
| 17 |
+
"eval_samples_per_second": 2033.424,
|
| 18 |
+
"eval_steps_per_second": 31.945,
|
| 19 |
+
"epoch": 5.0
|
| 20 |
+
}
|
| 21 |
+
}
|
split_test.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
split_train.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
split_valid.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|