Any-to-Any
Transformers
Safetensors
lfm3_audio
automatic-speech-recognition
liquid
lfm3
audio
speech
edge
conversational
multilingual
custom_code
Instructions to use eustlb/LFM3-Audio-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use eustlb/LFM3-Audio-2B with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModelForMultimodalLM model = AutoModelForMultimodalLM.from_pretrained("eustlb/LFM3-Audio-2B", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Rename the speech encoder sub-config to audio_config (lfm3_audio_audio)
Browse files- config.json +2 -2
config.json
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
"architectures": [
|
| 3 |
"Lfm3AudioForConditionalGeneration"
|
| 4 |
],
|
| 5 |
-
"
|
| 6 |
"attention_bias": false,
|
| 7 |
"attention_dropout": 0.0,
|
| 8 |
"conv_L_cache": 3,
|
|
@@ -32,7 +32,7 @@
|
|
| 32 |
"sliding_attention"
|
| 33 |
],
|
| 34 |
"max_position_embeddings": 128000,
|
| 35 |
-
"model_type": "
|
| 36 |
"norm_eps": 1e-05,
|
| 37 |
"num_attention_heads": 8,
|
| 38 |
"num_hidden_layers": 17,
|
|
|
|
| 2 |
"architectures": [
|
| 3 |
"Lfm3AudioForConditionalGeneration"
|
| 4 |
],
|
| 5 |
+
"audio_config": {
|
| 6 |
"attention_bias": false,
|
| 7 |
"attention_dropout": 0.0,
|
| 8 |
"conv_L_cache": 3,
|
|
|
|
| 32 |
"sliding_attention"
|
| 33 |
],
|
| 34 |
"max_position_embeddings": 128000,
|
| 35 |
+
"model_type": "lfm3_audio_audio",
|
| 36 |
"norm_eps": 1e-05,
|
| 37 |
"num_attention_heads": 8,
|
| 38 |
"num_hidden_layers": 17,
|