Upload 2 files
Browse files- config.json +34 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"VibeVoiceSemanticTokenizerModel"
|
| 4 |
+
],
|
| 5 |
+
"causal": true,
|
| 6 |
+
"channels": 1,
|
| 7 |
+
"conv_bias": true,
|
| 8 |
+
"conv_norm": "none",
|
| 9 |
+
"corpus_normalize": 0.0,
|
| 10 |
+
"disable_last_norm": true,
|
| 11 |
+
"encoder_depths": "3-3-3-3-3-3-8",
|
| 12 |
+
"encoder_n_filters": 32,
|
| 13 |
+
"encoder_ratios": [
|
| 14 |
+
8,
|
| 15 |
+
5,
|
| 16 |
+
5,
|
| 17 |
+
4,
|
| 18 |
+
2,
|
| 19 |
+
2
|
| 20 |
+
],
|
| 21 |
+
"fix_std": 0,
|
| 22 |
+
"layer_scale_init_value": 1e-06,
|
| 23 |
+
"layernorm": "RMSNorm",
|
| 24 |
+
"layernorm_elementwise_affine": true,
|
| 25 |
+
"layernorm_eps": 1e-05,
|
| 26 |
+
"mixer_layer": "depthwise_conv",
|
| 27 |
+
"model_type": "vibevoice_semantic_tokenizer",
|
| 28 |
+
"pad_mode": "constant",
|
| 29 |
+
"std_dist_type": "none",
|
| 30 |
+
"torch_dtype": "bfloat16",
|
| 31 |
+
"transformers_version": "4.51.3",
|
| 32 |
+
"vae_dim": 128,
|
| 33 |
+
"weight_init_value": 0.01
|
| 34 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a790160422970e11d793558424899e4b7bc89988341b2125170a930d1ce18fe4
|
| 3 |
+
size 689256688
|