upload weights
Browse files- encoder.json +18 -0
- encoder.safetensors +3 -0
- flowmodel.json +17 -0
- flowmodel.safetensors +3 -0
- gsdecoder.json +33 -0
- gsdecoder.safetensors +3 -0
- pipeline.json +96 -0
encoder.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "UniLatEncoder",
|
| 3 |
+
"args": {
|
| 4 |
+
"resolution": 64,
|
| 5 |
+
"in_channels": 1024,
|
| 6 |
+
"model_channels": 768,
|
| 7 |
+
"latent_channels": 8,
|
| 8 |
+
"num_blocks": 12,
|
| 9 |
+
"num_heads": 12,
|
| 10 |
+
"mlp_ratio": 4,
|
| 11 |
+
"attn_mode": "swin",
|
| 12 |
+
"window_size": 8,
|
| 13 |
+
"use_fp16": true,
|
| 14 |
+
"feats_latent_channels": 32,
|
| 15 |
+
"num_res_blocks": 4,
|
| 16 |
+
"pe_voxel_dim": 0
|
| 17 |
+
}
|
| 18 |
+
}
|
encoder.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1cad63b4d7c7611c9a4362d5f5ad784cf2e2be57f6c23a78670122df48f7ff22
|
| 3 |
+
size 703990224
|
flowmodel.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "SparseStructureFlowModel",
|
| 3 |
+
"args":{
|
| 4 |
+
"resolution": 16,
|
| 5 |
+
"in_channels": 32,
|
| 6 |
+
"out_channels": 32,
|
| 7 |
+
"model_channels": 1280,
|
| 8 |
+
"cond_channels": 1280,
|
| 9 |
+
"num_blocks": 36,
|
| 10 |
+
"num_heads": 32,
|
| 11 |
+
"mlp_ratio": 4,
|
| 12 |
+
"patch_size": 1,
|
| 13 |
+
"pe_mode": "ape",
|
| 14 |
+
"qk_rms_norm": true,
|
| 15 |
+
"use_fp16": true
|
| 16 |
+
}
|
| 17 |
+
}
|
flowmodel.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a4de14b31f8330d9fae38ff2cedb6f176e7d309944a401fbd402c84b8c57702c
|
| 3 |
+
size 5223945640
|
gsdecoder.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "UniLatGaussianDecoder",
|
| 3 |
+
"args": {
|
| 4 |
+
"resolution": 64,
|
| 5 |
+
"model_channels": 768,
|
| 6 |
+
"latent_channels": 8,
|
| 7 |
+
"num_blocks": 12,
|
| 8 |
+
"num_heads": 12,
|
| 9 |
+
"mlp_ratio": 4,
|
| 10 |
+
"attn_mode": "swin",
|
| 11 |
+
"window_size": 8,
|
| 12 |
+
"use_fp16": true,
|
| 13 |
+
"feats_latent_channels": 32,
|
| 14 |
+
"num_res_blocks": 4,
|
| 15 |
+
"representation_config": {
|
| 16 |
+
"lr": {
|
| 17 |
+
"_xyz": 1.0,
|
| 18 |
+
"_features_dc": 1.0,
|
| 19 |
+
"_opacity": 1.0,
|
| 20 |
+
"_scaling": 1.0,
|
| 21 |
+
"_rotation": 0.1
|
| 22 |
+
},
|
| 23 |
+
"perturb_offset": true,
|
| 24 |
+
"voxel_size": 1.5,
|
| 25 |
+
"num_gaussians": 32,
|
| 26 |
+
"2d_filter_kernel_size": 0.1,
|
| 27 |
+
"3d_filter_kernel_size": 0.0009,
|
| 28 |
+
"scaling_bias": 0.004,
|
| 29 |
+
"opacity_bias": 0.1,
|
| 30 |
+
"scaling_activation": "softplus"
|
| 31 |
+
}
|
| 32 |
+
}
|
| 33 |
+
}
|
gsdecoder.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b61852caae688fc0acf5a7721b84856227c977239dc9daa33e156632051f941
|
| 3 |
+
size 1054355788
|
pipeline.json
ADDED
|
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "UniLat3DImageTo3DPipeline",
|
| 3 |
+
"args":{
|
| 4 |
+
"models":{
|
| 5 |
+
"sparse_structure_flow_model": "flowmodel",
|
| 6 |
+
"unilat_decoder_gs": "gsdecoder"
|
| 7 |
+
},
|
| 8 |
+
"unilat_sampler":{
|
| 9 |
+
"name": "FlowEulerGuidanceIntervalSampler",
|
| 10 |
+
"args": {
|
| 11 |
+
"sigma_min": 1e-05
|
| 12 |
+
},
|
| 13 |
+
"params": {
|
| 14 |
+
"steps": 35,
|
| 15 |
+
"cfg_strength": 5.0,
|
| 16 |
+
"cfg_interval": [
|
| 17 |
+
0.5,
|
| 18 |
+
1.0
|
| 19 |
+
],
|
| 20 |
+
"rescale_t": 3.0
|
| 21 |
+
}
|
| 22 |
+
},
|
| 23 |
+
"image_cond_model": "dinov3_vith16plus",
|
| 24 |
+
"unilat_normalization": {
|
| 25 |
+
"mean": [
|
| 26 |
+
0.046790410882579354,
|
| 27 |
+
-0.029926393560248885,
|
| 28 |
+
-0.016157663277648754,
|
| 29 |
+
-0.03144306153190407,
|
| 30 |
+
-0.02231043000268882,
|
| 31 |
+
-0.04343507226115912,
|
| 32 |
+
0.014098768188985452,
|
| 33 |
+
0.014915844535938036,
|
| 34 |
+
-0.05923355591278892,
|
| 35 |
+
-0.040837326425598654,
|
| 36 |
+
-0.11795711425943962,
|
| 37 |
+
-0.0009295193267654697,
|
| 38 |
+
-0.04317192372130144,
|
| 39 |
+
0.056539989003499797,
|
| 40 |
+
0.0021068858674523957,
|
| 41 |
+
-0.02188881225342723,
|
| 42 |
+
-0.002516907216348253,
|
| 43 |
+
-0.03481384488869907,
|
| 44 |
+
-0.08271037548757688,
|
| 45 |
+
-0.04577478785797573,
|
| 46 |
+
-0.06576960891947232,
|
| 47 |
+
0.13284408929184366,
|
| 48 |
+
-0.00443547622894399,
|
| 49 |
+
-0.047641309242299736,
|
| 50 |
+
0.03903243565243531,
|
| 51 |
+
0.026947632154379193,
|
| 52 |
+
-0.031267948436940925,
|
| 53 |
+
0.054134701503369374,
|
| 54 |
+
0.05544821323388031,
|
| 55 |
+
-0.0005631704853942428,
|
| 56 |
+
-0.008881308470383322,
|
| 57 |
+
0.05028240915789243
|
| 58 |
+
],
|
| 59 |
+
"std": [
|
| 60 |
+
0.4689496681692278,
|
| 61 |
+
0.37089456130040843,
|
| 62 |
+
0.3465033475496703,
|
| 63 |
+
0.37363125050291796,
|
| 64 |
+
0.41707908196635946,
|
| 65 |
+
0.5187476970672488,
|
| 66 |
+
0.4922019643514823,
|
| 67 |
+
0.4016635057709841,
|
| 68 |
+
0.5075262968414974,
|
| 69 |
+
0.44471242290718616,
|
| 70 |
+
0.4524994324372127,
|
| 71 |
+
0.5297438570844593,
|
| 72 |
+
0.39098343177898726,
|
| 73 |
+
0.3701102372843775,
|
| 74 |
+
0.3635129436071737,
|
| 75 |
+
0.5119164495168363,
|
| 76 |
+
1.048898670302766,
|
| 77 |
+
0.6799519623331088,
|
| 78 |
+
0.41942227875125576,
|
| 79 |
+
0.4853593390342862,
|
| 80 |
+
0.48926874411154825,
|
| 81 |
+
0.4897713630282384,
|
| 82 |
+
0.37380177939311887,
|
| 83 |
+
0.465606865826469,
|
| 84 |
+
0.34167164099303804,
|
| 85 |
+
0.3796393262247079,
|
| 86 |
+
0.6043801640252177,
|
| 87 |
+
0.4688239908160894,
|
| 88 |
+
0.5275023072468689,
|
| 89 |
+
0.3905801437065808,
|
| 90 |
+
0.430450528949646,
|
| 91 |
+
0.35630311522795705
|
| 92 |
+
]
|
| 93 |
+
}
|
| 94 |
+
|
| 95 |
+
}
|
| 96 |
+
}
|