Config: projector_hidden_act, mixed_end_token_id; processor: live transcription chunk sizes
Browse files- config.json +2 -0
- processor_config.json +3 -1
config.json
CHANGED
|
@@ -102,9 +102,11 @@
|
|
| 102 |
4,
|
| 103 |
4
|
| 104 |
],
|
|
|
|
| 105 |
"mixed_start_token_id": 125015,
|
| 106 |
"model_type": "lfm3_audio",
|
| 107 |
"num_codebooks": 12,
|
|
|
|
| 108 |
"text_config": {
|
| 109 |
"attention_bias": false,
|
| 110 |
"attention_dropout": 0.0,
|
|
|
|
| 102 |
4,
|
| 103 |
4
|
| 104 |
],
|
| 105 |
+
"mixed_end_token_id": 125016,
|
| 106 |
"mixed_start_token_id": 125015,
|
| 107 |
"model_type": "lfm3_audio",
|
| 108 |
"num_codebooks": 12,
|
| 109 |
+
"projector_hidden_act": "gelu",
|
| 110 |
"text_config": {
|
| 111 |
"attention_bias": false,
|
| 112 |
"attention_dropout": 0.0,
|
processor_config.json
CHANGED
|
@@ -20,5 +20,7 @@
|
|
| 20 |
"sampling_rate": 16000,
|
| 21 |
"win_length": 400
|
| 22 |
},
|
| 23 |
-
"processor_class": "Lfm3AudioProcessor"
|
|
|
|
|
|
|
| 24 |
}
|
|
|
|
| 20 |
"sampling_rate": 16000,
|
| 21 |
"win_length": 400
|
| 22 |
},
|
| 23 |
+
"processor_class": "Lfm3AudioProcessor",
|
| 24 |
+
"num_audio_tokens_per_chunk": 12,
|
| 25 |
+
"max_audio_chunks_per_context": 31
|
| 26 |
}
|