Upload folder using huggingface_hub
Browse files- config.json +44 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "sam_audio",
|
| 3 |
+
"model_size": "base",
|
| 4 |
+
"in_channels": 768,
|
| 5 |
+
"audio_codec": {
|
| 6 |
+
"encoder_dim": 64,
|
| 7 |
+
"encoder_rates": [
|
| 8 |
+
2,
|
| 9 |
+
8,
|
| 10 |
+
10,
|
| 11 |
+
12
|
| 12 |
+
],
|
| 13 |
+
"latent_dim": 1024,
|
| 14 |
+
"decoder_dim": 1536,
|
| 15 |
+
"decoder_rates": [
|
| 16 |
+
12,
|
| 17 |
+
10,
|
| 18 |
+
8,
|
| 19 |
+
2
|
| 20 |
+
],
|
| 21 |
+
"n_codebooks": 16,
|
| 22 |
+
"codebook_size": 1024,
|
| 23 |
+
"codebook_dim": 128,
|
| 24 |
+
"sample_rate": 48000
|
| 25 |
+
},
|
| 26 |
+
"text_encoder": {
|
| 27 |
+
"name": "t5-base",
|
| 28 |
+
"max_length": 512,
|
| 29 |
+
"dim": 768
|
| 30 |
+
},
|
| 31 |
+
"transformer": {
|
| 32 |
+
"dim": 2048,
|
| 33 |
+
"n_heads": 16,
|
| 34 |
+
"n_layers": 16,
|
| 35 |
+
"dropout": 0.1,
|
| 36 |
+
"qk_norm": true,
|
| 37 |
+
"fc_bias": false,
|
| 38 |
+
"ffn_exp": 4,
|
| 39 |
+
"context_dim": 2048,
|
| 40 |
+
"out_channels": 256
|
| 41 |
+
},
|
| 42 |
+
"num_anchors": 3,
|
| 43 |
+
"anchor_embedding_dim": 128
|
| 44 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7fa64c16b4e72f4486b05fa7993edc2a40fd13891089495b3e5e0be36b3232b
|
| 3 |
+
size 5032052232
|