Automatic Speech Recognition
Transformers
Safetensors
phoneticxeus
feature-extraction
phone-recognition
ipa
ctc
multilingual
xeus
custom_code
Instructions to use changelinglab/PhoneticXeus with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use changelinglab/PhoneticXeus with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="changelinglab/PhoneticXeus", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("changelinglab/PhoneticXeus", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download config_tree.log from changelinglab/PhoneticXeus: direct link, hf CLI and curl.
- Browser
- Download file 11.5 kB
-
https://huggingface.co/changelinglab/PhoneticXeus/resolve/main/config_tree.log
- Command line
-
hf download hf://changelinglab/PhoneticXeus/config_tree.log
-
curl -L -o config_tree.log https://huggingface.co/changelinglab/PhoneticXeus/resolve/main/config_tree.log
11.5 kB
| CONFIG | |
| βββ data | |
| β βββ _target_: src.data.kaldi_pretraining_dataset.build_kaldi_datamodule | |
| β train_split: train | |
| β dev_splits: | |
| β - dev_1k | |
| β - dev_gmuaccent | |
| β - dev_buckeye | |
| β - dev_epadb | |
| β - dev_speechoceanotth | |
| β - dev_l2arctic | |
| β dataset_config_path: configs/data/ipapack_index.yaml | |
| β vocab_file: src/model/xeusphoneme/resources/ipa_vocab.json | |
| β batch_size: 12 | |
| β num_workers: 6 | |
| β | |
| βββ model | |
| β βββ _target_: src.recipe.phone_recognition.model_module.PhoneRecognitionMode | |
| β optimizer: | |
| β _target_: torch.optim.Adam | |
| β _partial_: true | |
| β lr: 3.0e-05 | |
| β weight_decay: 0.0 | |
| β scheduler: | |
| β _target_: src.recipe.phone_recognition.model_module.get_w2v2ph_schedul | |
| β _partial_: true | |
| β num_training_steps: 40000 | |
| β encoder_unfreeze_step: 3000 | |
| β warmup_fraction: 0.05 | |
| β constant_fraction: 0.75 | |
| β net: | |
| β _target_: src.model.xeusphoneme.builders.build_xeus_pr_from_hf | |
| β work_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp/cache/xeus | |
| β hf_repo: espnet/xeus | |
| β force: false | |
| β config_file: null | |
| β checkpoint: null | |
| β vocab_file: src/model/xeusphoneme/resources/ipa_vocab.json | |
| β load_ckpt: true | |
| β interctc_weight: 0.3 | |
| β interctc_layer_idx: | |
| β - 4 | |
| β - 8 | |
| β - 12 | |
| β interctc_use_conditioning: true | |
| β ctc_weight: 1.0 | |
| β decoder_config: null | |
| β ctc_config: | |
| β ctc_type: builtin | |
| β dev_splits: | |
| β - dev_1k | |
| β - dev_gmuaccent | |
| β - dev_buckeye | |
| β - dev_epadb | |
| β - dev_speechoceanotth | |
| β - dev_l2arctic | |
| β | |
| βββ callbacks | |
| β βββ model_checkpoint: | |
| β _target_: lightning.pytorch.callbacks.ModelCheckpoint | |
| β dirpath: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp/runs/ipaaccent_ | |
| β filename: checkpoint-{step} | |
| β monitor: null | |
| β verbose: false | |
| β save_last: true | |
| β save_top_k: -1 | |
| β mode: min | |
| β auto_insert_metric_name: false | |
| β save_weights_only: false | |
| β every_n_train_steps: 2000 | |
| β train_time_interval: null | |
| β every_n_epochs: null | |
| β save_on_train_epoch_end: null | |
| β learning_rate_monitor: | |
| β _target_: lightning.pytorch.callbacks.LearningRateMonitor | |
| β logging_interval: step | |
| β early_stopping: null | |
| β model_summary: | |
| β _target_: lightning.pytorch.callbacks.RichModelSummary | |
| β max_depth: 1 | |
| β rich_progress_bar: | |
| β _target_: lightning.pytorch.callbacks.RichProgressBar | |
| β | |
| βββ logger | |
| β βββ wandb: | |
| β _target_: lightning.pytorch.loggers.wandb.WandbLogger | |
| β save_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp/runs/ipaaccent | |
| β offline: false | |
| β id: null | |
| β anonymous: null | |
| β project: SSLPR | |
| β log_model: false | |
| β prefix: '' | |
| β group: ipaaccent_ctc | |
| β tags: | |
| β - ipapack_accent | |
| β - ctc | |
| β - phone_recognition | |
| β job_type: '' | |
| β name: xeus_multiaccent.selfctc_l4_8_12.bs256.lr3em5.sched_p05warm_p75c | |
| β | |
| βββ trainer | |
| β βββ _target_: lightning.pytorch.trainer.Trainer | |
| β default_root_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp/runs/ipa | |
| β min_epochs: 0 | |
| β max_epochs: -1 | |
| β accelerator: gpu | |
| β devices: auto | |
| β check_val_every_n_epoch: null | |
| β deterministic: false | |
| β num_nodes: 1 | |
| β strategy: ddp_find_unused_parameters_true | |
| β precision: bf16 | |
| β sync_batchnorm: true | |
| β max_steps: 40000 | |
| β gradient_clip_val: 1.0 | |
| β accumulate_grad_batches: 5 | |
| β val_check_interval: 2000 | |
| β log_every_n_steps: 200 | |
| β | |
| βββ paths | |
| β βββ root_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench | |
| β exp_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp | |
| β download_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp/download | |
| β cache_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp/cache | |
| β output_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench/exp/runs/ipaaccent | |
| β work_dir: /mnt/home/chinjoul-andr-daf644/PhoneBench | |
| β | |
| βββ extras | |
| β βββ ignore_warnings: false | |
| β enforce_tags: true | |
| β print_config: true | |
| β | |
| βββ task_name | |
| β βββ ipaaccent_ctc | |
| βββ tags | |
| β βββ ['ipapack_accent', 'ctc', 'phone_recognition'] | |
| βββ train | |
| β βββ True | |
| βββ test | |
| β βββ True | |
| βββ predict | |
| β βββ False | |
| βββ pred_file | |
| β βββ None | |
| βββ run_folder | |
| β βββ xeus_multiaccent.selfctc_l4_8_12.bs256.lr3em5.sched_p05warm_p75const_3ku | |
| βββ ckpt_path | |
| β βββ None | |
| βββ seed | |
| βββ 42 | |