Upload diffusion_decoder/config.json with huggingface_hub
Browse files
diffusion_decoder/config.json
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_class_name": "LTX2VideoDiffusionDecoderModel",
|
| 3 |
+
"_diffusers_version": "0.40.0.dev0",
|
| 4 |
+
"decoder_head_dim": 64,
|
| 5 |
+
"decoder_model_output_type": "x0",
|
| 6 |
+
"decoder_num_inference_steps": 1,
|
| 7 |
+
"decoder_stage5_kernel": [
|
| 8 |
+
11,
|
| 9 |
+
11,
|
| 10 |
+
11
|
| 11 |
+
],
|
| 12 |
+
"decoder_stage_channels": [
|
| 13 |
+
2048,
|
| 14 |
+
1024,
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
256
|
| 18 |
+
],
|
| 19 |
+
"decoder_stage_depths": [
|
| 20 |
+
4,
|
| 21 |
+
6,
|
| 22 |
+
4,
|
| 23 |
+
2,
|
| 24 |
+
8
|
| 25 |
+
],
|
| 26 |
+
"decoder_stage_kernels": [
|
| 27 |
+
[
|
| 28 |
+
3,
|
| 29 |
+
7,
|
| 30 |
+
7
|
| 31 |
+
],
|
| 32 |
+
[
|
| 33 |
+
3,
|
| 34 |
+
7,
|
| 35 |
+
7
|
| 36 |
+
],
|
| 37 |
+
[
|
| 38 |
+
3,
|
| 39 |
+
5,
|
| 40 |
+
5
|
| 41 |
+
],
|
| 42 |
+
[
|
| 43 |
+
3,
|
| 44 |
+
5,
|
| 45 |
+
5
|
| 46 |
+
]
|
| 47 |
+
],
|
| 48 |
+
"decoder_t_emb_dim": 384,
|
| 49 |
+
"decoder_timestep_scale_multiplier": 1000.0,
|
| 50 |
+
"decoder_upsample_channel_reductions": [
|
| 51 |
+
2,
|
| 52 |
+
2,
|
| 53 |
+
1,
|
| 54 |
+
2
|
| 55 |
+
],
|
| 56 |
+
"decoder_upsample_strides": [
|
| 57 |
+
[
|
| 58 |
+
1,
|
| 59 |
+
2,
|
| 60 |
+
2
|
| 61 |
+
],
|
| 62 |
+
[
|
| 63 |
+
2,
|
| 64 |
+
1,
|
| 65 |
+
1
|
| 66 |
+
],
|
| 67 |
+
[
|
| 68 |
+
2,
|
| 69 |
+
2,
|
| 70 |
+
2
|
| 71 |
+
],
|
| 72 |
+
[
|
| 73 |
+
2,
|
| 74 |
+
2,
|
| 75 |
+
2
|
| 76 |
+
]
|
| 77 |
+
],
|
| 78 |
+
"latent_channels": 128,
|
| 79 |
+
"out_channels": 3,
|
| 80 |
+
"patch_size": 4,
|
| 81 |
+
"scaling_factor": 1.0,
|
| 82 |
+
"spatial_compression_ratio": 32,
|
| 83 |
+
"temporal_compression_ratio": 8
|
| 84 |
+
}
|