Mirror stabilityai/SAME-S → same-s/
Browse files- same-s/model.safetensors +3 -0
- same-s/model_config.json +70 -0
same-s/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c19698ce3a0b462acb967ee495e9eb7945221f236968c50206cce8cf22b3d305
|
| 3 |
+
size 433007588
|
same-s/model_config.json
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "autoencoder",
|
| 3 |
+
"sample_size": 24576,
|
| 4 |
+
"sample_rate": 44100,
|
| 5 |
+
"audio_channels": 2,
|
| 6 |
+
"model": {
|
| 7 |
+
"pretransform": {
|
| 8 |
+
"type": "patched",
|
| 9 |
+
"config": {
|
| 10 |
+
"patch_size": 256,
|
| 11 |
+
"channels": 2
|
| 12 |
+
}
|
| 13 |
+
},
|
| 14 |
+
"encoder": {
|
| 15 |
+
"type": "same",
|
| 16 |
+
"requires_grad": false,
|
| 17 |
+
"config": {
|
| 18 |
+
"in_channels": 512,
|
| 19 |
+
"channels": 128,
|
| 20 |
+
"c_mults": [6],
|
| 21 |
+
"strides": [16],
|
| 22 |
+
"latent_dim": 256,
|
| 23 |
+
"transformer_depths": [6],
|
| 24 |
+
"checkpointing": false,
|
| 25 |
+
"differential": true,
|
| 26 |
+
"dyt": true,
|
| 27 |
+
"dim_heads": 64,
|
| 28 |
+
"variable_stride": true,
|
| 29 |
+
"chunk_size": 32,
|
| 30 |
+
"chunk_midpoint_shift": true,
|
| 31 |
+
"mask_noise": 0.0
|
| 32 |
+
}
|
| 33 |
+
},
|
| 34 |
+
"decoder": {
|
| 35 |
+
"type": "same",
|
| 36 |
+
"requires_grad": false,
|
| 37 |
+
"config": {
|
| 38 |
+
"out_channels": 512,
|
| 39 |
+
"channels": 128,
|
| 40 |
+
"c_mults": [6],
|
| 41 |
+
"strides": [16],
|
| 42 |
+
"latent_dim": 256,
|
| 43 |
+
"transformer_depths": [6],
|
| 44 |
+
"sinusoidal_blocks": [0],
|
| 45 |
+
"checkpointing": false,
|
| 46 |
+
"differential": true,
|
| 47 |
+
"dyt": true,
|
| 48 |
+
"dim_heads": 64,
|
| 49 |
+
"variable_stride": true,
|
| 50 |
+
"chunk_size": 32,
|
| 51 |
+
"chunk_midpoint_shift": true,
|
| 52 |
+
"conv_mapping": true,
|
| 53 |
+
"mask_noise": 0.01
|
| 54 |
+
}
|
| 55 |
+
},
|
| 56 |
+
"bottleneck": {
|
| 57 |
+
"type": "softnorm",
|
| 58 |
+
"config": {
|
| 59 |
+
"dim": 256,
|
| 60 |
+
"noise_augment_dim": 0,
|
| 61 |
+
"noise_regularize": true,
|
| 62 |
+
"auto_scale": true,
|
| 63 |
+
"freeze": true
|
| 64 |
+
}
|
| 65 |
+
},
|
| 66 |
+
"latent_dim": 256,
|
| 67 |
+
"downsampling_ratio": 4096,
|
| 68 |
+
"io_channels": 2
|
| 69 |
+
}
|
| 70 |
+
}
|