LanguageBind
commited on
Commit
•
2f4d0b8
1
Parent(s):
a1be3f7
Upload 2 files
Browse files- checkpoint.ckpt +3 -0
- config.json +85 -0
checkpoint.ckpt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a7123138121651f316e14d6c6337f1f0b45b8c7b533f129da95e74219c0900c2
|
3 |
+
size 956837970
|
config.json
ADDED
@@ -0,0 +1,85 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "CausalVAEModel",
|
3 |
+
"_diffusers_version": "0.27.2",
|
4 |
+
"attn_resolutions": [],
|
5 |
+
"decoder_attention": "AttnBlock3DFix",
|
6 |
+
"decoder_conv_in": "CausalConv3d",
|
7 |
+
"decoder_conv_out": "CausalConv3d",
|
8 |
+
"decoder_mid_resnet": "ResnetBlock3D",
|
9 |
+
"decoder_resnet_blocks": [
|
10 |
+
"ResnetBlock3D",
|
11 |
+
"ResnetBlock3D",
|
12 |
+
"ResnetBlock3D",
|
13 |
+
"ResnetBlock3D"
|
14 |
+
],
|
15 |
+
"decoder_spatial_upsample": [
|
16 |
+
"",
|
17 |
+
"SpatialUpsample2x",
|
18 |
+
"Spatial2xTime2x3DUpsample",
|
19 |
+
"Spatial2xTime2x3DUpsample"
|
20 |
+
],
|
21 |
+
"decoder_spatial_upsample_unup": [
|
22 |
+
"",
|
23 |
+
"",
|
24 |
+
"",
|
25 |
+
""
|
26 |
+
],
|
27 |
+
"decoder_temporal_upsample": [
|
28 |
+
"",
|
29 |
+
"",
|
30 |
+
"",
|
31 |
+
""
|
32 |
+
],
|
33 |
+
"double_z": true,
|
34 |
+
"dropout": 0.0,
|
35 |
+
"embed_dim": 4,
|
36 |
+
"encoder_attention": "AttnBlock3DFix",
|
37 |
+
"encoder_conv_in": "Conv2d",
|
38 |
+
"encoder_conv_out": "CausalConv3d",
|
39 |
+
"encoder_mid_resnet": "ResnetBlock3D",
|
40 |
+
"encoder_resnet_blocks": [
|
41 |
+
"ResnetBlock2D",
|
42 |
+
"ResnetBlock2D",
|
43 |
+
"ResnetBlock3D",
|
44 |
+
"ResnetBlock3D"
|
45 |
+
],
|
46 |
+
"encoder_spatial_downsample": [
|
47 |
+
"Downsample",
|
48 |
+
"Spatial2xTime2x3DDownsample",
|
49 |
+
"Spatial2xTime2x3DDownsample",
|
50 |
+
""
|
51 |
+
],
|
52 |
+
"encoder_spatial_downsample_undown": [
|
53 |
+
"",
|
54 |
+
"",
|
55 |
+
"",
|
56 |
+
""
|
57 |
+
],
|
58 |
+
"encoder_temporal_downsample": [
|
59 |
+
"",
|
60 |
+
"",
|
61 |
+
"",
|
62 |
+
""
|
63 |
+
],
|
64 |
+
"hidden_size": 128,
|
65 |
+
"hidden_size_mult": [
|
66 |
+
1,
|
67 |
+
2,
|
68 |
+
4,
|
69 |
+
4
|
70 |
+
],
|
71 |
+
"in_channels": 3,
|
72 |
+
"loss_params": {
|
73 |
+
"disc_start": 1,
|
74 |
+
"disc_weight": 0.5,
|
75 |
+
"kl_weight": 1e-06,
|
76 |
+
"logvar_init": 0.0
|
77 |
+
},
|
78 |
+
"loss_type": "opensora.models.ae.videobase.losses.LPIPSWithDiscriminator",
|
79 |
+
"lr": 1e-05,
|
80 |
+
"num_res_blocks": 2,
|
81 |
+
"out_channels": 3,
|
82 |
+
"q_conv": "CausalConv3d",
|
83 |
+
"resolution": 256,
|
84 |
+
"z_channels": 4
|
85 |
+
}
|