upload vae and transformer
Browse files
transformer/config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_class_name": "WanTransformer3DModel",
|
| 3 |
+
"_diffusers_version": "0.35.2",
|
| 4 |
+
"added_kv_proj_dim": null,
|
| 5 |
+
"attention_head_dim": 128,
|
| 6 |
+
"attention_type": "sla",
|
| 7 |
+
"cross_attn_norm": true,
|
| 8 |
+
"eps": 1e-06,
|
| 9 |
+
"ffn_dim": 8960,
|
| 10 |
+
"freq_dim": 256,
|
| 11 |
+
"image_dim": null,
|
| 12 |
+
"in_channels": 16,
|
| 13 |
+
"num_attention_heads": 12,
|
| 14 |
+
"num_layers": 30,
|
| 15 |
+
"out_channels": 16,
|
| 16 |
+
"patch_size": [
|
| 17 |
+
1,
|
| 18 |
+
2,
|
| 19 |
+
2
|
| 20 |
+
],
|
| 21 |
+
"pos_embed_seq_len": null,
|
| 22 |
+
"qk_norm": "rms_norm_across_heads",
|
| 23 |
+
"rope_max_seq_len": 1024,
|
| 24 |
+
"sla_topk": 0.1,
|
| 25 |
+
"text_dim": 4096
|
| 26 |
+
}
|
transformer/diffusion_pytorch_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3144ec25ab4c84f79e42f5bf08d7c7141e60e9665a6f76661853e3d8c3b20215
|
| 3 |
+
size 2839627208
|
vae/config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_class_name": "AutoencoderKLWan",
|
| 3 |
+
"_diffusers_version": "0.35.2",
|
| 4 |
+
"attn_scales": [],
|
| 5 |
+
"base_dim": 96,
|
| 6 |
+
"decoder_base_dim": null,
|
| 7 |
+
"dim_mult": [
|
| 8 |
+
1,
|
| 9 |
+
2,
|
| 10 |
+
4,
|
| 11 |
+
4
|
| 12 |
+
],
|
| 13 |
+
"dropout": 0.0,
|
| 14 |
+
"in_channels": 3,
|
| 15 |
+
"is_residual": false,
|
| 16 |
+
"latents_mean": [
|
| 17 |
+
-0.7571,
|
| 18 |
+
-0.7089,
|
| 19 |
+
-0.9113,
|
| 20 |
+
0.1075,
|
| 21 |
+
-0.1745,
|
| 22 |
+
0.9653,
|
| 23 |
+
-0.1517,
|
| 24 |
+
1.5508,
|
| 25 |
+
0.4134,
|
| 26 |
+
-0.0715,
|
| 27 |
+
0.5517,
|
| 28 |
+
-0.3632,
|
| 29 |
+
-0.1922,
|
| 30 |
+
-0.9497,
|
| 31 |
+
0.2503,
|
| 32 |
+
-0.2921
|
| 33 |
+
],
|
| 34 |
+
"latents_std": [
|
| 35 |
+
2.8184,
|
| 36 |
+
1.4541,
|
| 37 |
+
2.3275,
|
| 38 |
+
2.6558,
|
| 39 |
+
1.2196,
|
| 40 |
+
1.7708,
|
| 41 |
+
2.6052,
|
| 42 |
+
2.0743,
|
| 43 |
+
3.2687,
|
| 44 |
+
2.1526,
|
| 45 |
+
2.8652,
|
| 46 |
+
1.5579,
|
| 47 |
+
1.6382,
|
| 48 |
+
1.1253,
|
| 49 |
+
2.8251,
|
| 50 |
+
1.916
|
| 51 |
+
],
|
| 52 |
+
"num_res_blocks": 2,
|
| 53 |
+
"out_channels": 3,
|
| 54 |
+
"patch_size": null,
|
| 55 |
+
"scale_factor_spatial": 8,
|
| 56 |
+
"scale_factor_temporal": 4,
|
| 57 |
+
"temperal_downsample": [
|
| 58 |
+
false,
|
| 59 |
+
true,
|
| 60 |
+
true
|
| 61 |
+
],
|
| 62 |
+
"z_dim": 16
|
| 63 |
+
}
|
vae/diffusion_pytorch_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d6e524b3fffede1787a74e81b30976dce5400c4439ba64222168e607ed19e793
|
| 3 |
+
size 507591892
|