Instructions to use onnx-community/Voxtral-Mini-3B-2507-ONNX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use onnx-community/Voxtral-Mini-3B-2507-ONNX with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('audio-text-to-text', 'onnx-community/Voxtral-Mini-3B-2507-ONNX');
Update config.json
Browse files- config.json +49 -25
config.json
CHANGED
|
@@ -1,32 +1,55 @@
|
|
| 1 |
{
|
| 2 |
"architectures": [
|
| 3 |
-
"
|
| 4 |
],
|
| 5 |
-
"
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
"hidden_size": 3072,
|
| 12 |
-
"
|
| 13 |
-
"
|
| 14 |
-
"
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
"transformers_version": "4.54.0.dev0",
|
| 28 |
-
|
| 29 |
-
"vocab_size": 131072,
|
| 30 |
"transformers.js_config": {
|
| 31 |
"dtype": {
|
| 32 |
"embed_tokens": "fp16"
|
|
@@ -43,5 +66,6 @@
|
|
| 43 |
"embed_tokens.onnx": 1,
|
| 44 |
"embed_tokens_fp16.onnx": 1
|
| 45 |
}
|
| 46 |
-
}
|
|
|
|
| 47 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"architectures": [
|
| 3 |
+
"VoxtralForConditionalGeneration"
|
| 4 |
],
|
| 5 |
+
"audio_config": {
|
| 6 |
+
"activation_dropout": 0.0,
|
| 7 |
+
"activation_function": "gelu",
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"dropout": 0.0,
|
| 10 |
+
"head_dim": 64,
|
| 11 |
+
"hidden_size": 1280,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 5120,
|
| 14 |
+
"layerdrop": 0.0,
|
| 15 |
+
"max_source_positions": 1500,
|
| 16 |
+
"model_type": "voxtral_encoder",
|
| 17 |
+
"num_attention_heads": 20,
|
| 18 |
+
"num_hidden_layers": 32,
|
| 19 |
+
"num_key_value_heads": 20,
|
| 20 |
+
"num_mel_bins": 128,
|
| 21 |
+
"scale_embedding": false,
|
| 22 |
+
"vocab_size": 51866
|
| 23 |
+
},
|
| 24 |
+
"audio_token_id": 24,
|
| 25 |
"hidden_size": 3072,
|
| 26 |
+
"model_type": "voxtral",
|
| 27 |
+
"projector_hidden_act": "gelu",
|
| 28 |
+
"text_config": {
|
| 29 |
+
"attention_bias": false,
|
| 30 |
+
"attention_dropout": 0.0,
|
| 31 |
+
"head_dim": 128,
|
| 32 |
+
"hidden_act": "silu",
|
| 33 |
+
"hidden_size": 3072,
|
| 34 |
+
"initializer_range": 0.02,
|
| 35 |
+
"intermediate_size": 8192,
|
| 36 |
+
"max_position_embeddings": 131072,
|
| 37 |
+
"mlp_bias": false,
|
| 38 |
+
"model_type": "llama",
|
| 39 |
+
"num_attention_heads": 32,
|
| 40 |
+
"num_hidden_layers": 30,
|
| 41 |
+
"num_key_value_heads": 8,
|
| 42 |
+
"pretraining_tp": 1,
|
| 43 |
+
"rms_norm_eps": 1e-05,
|
| 44 |
+
"rope_scaling": null,
|
| 45 |
+
"rope_theta": 100000000.0,
|
| 46 |
+
"sliding_window": null,
|
| 47 |
+
"use_cache": true,
|
| 48 |
+
"vocab_size": 131072
|
| 49 |
+
},
|
| 50 |
+
"torch_dtype": "bfloat16",
|
| 51 |
"transformers_version": "4.54.0.dev0",
|
| 52 |
+
|
|
|
|
| 53 |
"transformers.js_config": {
|
| 54 |
"dtype": {
|
| 55 |
"embed_tokens": "fp16"
|
|
|
|
| 66 |
"embed_tokens.onnx": 1,
|
| 67 |
"embed_tokens_fp16.onnx": 1
|
| 68 |
}
|
| 69 |
+
},
|
| 70 |
+
"vocab_size": 131072
|
| 71 |
}
|