Upload folder using huggingface_hub
Browse files- ckpts/shape_slat_cond_projection.safetensors +3 -0
- ckpts/shape_slat_denoiser.json +18 -0
- ckpts/shape_slat_denoiser.safetensors +3 -0
- ckpts/sparse_structure_cond_projection.safetensors +3 -0
- ckpts/sparse_structure_denoiser.json +18 -0
- ckpts/sparse_structure_denoiser.safetensors +3 -0
- ckpts/tex_slat_cond_projection.safetensors +3 -0
- ckpts/tex_slat_denoiser.json +18 -0
- ckpts/tex_slat_denoiser.safetensors +3 -0
- pipeline.json +131 -0
ckpts/shape_slat_cond_projection.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ff56e98c0477c05bc18e9e161dbc8c20627913b150f576f7f98aa9ef483b776
|
| 3 |
+
size 8401216
|
ckpts/shape_slat_denoiser.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "ElasticSLatFlowModel",
|
| 3 |
+
"args": {
|
| 4 |
+
"resolution": 32,
|
| 5 |
+
"in_channels": 32,
|
| 6 |
+
"out_channels": 32,
|
| 7 |
+
"model_channels": 1536,
|
| 8 |
+
"cond_channels": 1024,
|
| 9 |
+
"num_blocks": 30,
|
| 10 |
+
"num_heads": 12,
|
| 11 |
+
"mlp_ratio": 5.3334,
|
| 12 |
+
"pe_mode": "rope",
|
| 13 |
+
"share_mod": true,
|
| 14 |
+
"initialization": "scaled",
|
| 15 |
+
"qk_rms_norm": true,
|
| 16 |
+
"qk_rms_norm_cross": true
|
| 17 |
+
}
|
| 18 |
+
}
|
ckpts/shape_slat_denoiser.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dabc34f29546fdd15ed69ec27672d3bccd4550ac9af212a477df3ed001a3300b
|
| 3 |
+
size 5169081512
|
ckpts/sparse_structure_cond_projection.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33074d9cfa7e0b14f93789863850e6ade5d30170e9591a4c85987e1c41656892
|
| 3 |
+
size 8401216
|
ckpts/sparse_structure_denoiser.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "SparseStructureFlowModel",
|
| 3 |
+
"args": {
|
| 4 |
+
"resolution": 16,
|
| 5 |
+
"in_channels": 8,
|
| 6 |
+
"out_channels": 8,
|
| 7 |
+
"model_channels": 1536,
|
| 8 |
+
"cond_channels": 1024,
|
| 9 |
+
"num_blocks": 30,
|
| 10 |
+
"num_heads": 12,
|
| 11 |
+
"mlp_ratio": 5.3334,
|
| 12 |
+
"pe_mode": "rope",
|
| 13 |
+
"share_mod": true,
|
| 14 |
+
"initialization": "scaled",
|
| 15 |
+
"qk_rms_norm": true,
|
| 16 |
+
"qk_rms_norm_cross": true
|
| 17 |
+
}
|
| 18 |
+
}
|
ckpts/sparse_structure_denoiser.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:897e6e2b5017a89963058adbf3e6321aac795ff5a3791c04e584d2db2a556281
|
| 3 |
+
size 5170883736
|
ckpts/tex_slat_cond_projection.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f1063541f713d0ff34c0ccf20cb59b97c762ed7a90a2c4345ab1e893677c3bd
|
| 3 |
+
size 8401216
|
ckpts/tex_slat_denoiser.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "ElasticSLatFlowModel",
|
| 3 |
+
"args": {
|
| 4 |
+
"resolution": 32,
|
| 5 |
+
"in_channels": 64,
|
| 6 |
+
"out_channels": 32,
|
| 7 |
+
"model_channels": 1536,
|
| 8 |
+
"cond_channels": 1024,
|
| 9 |
+
"num_blocks": 30,
|
| 10 |
+
"num_heads": 12,
|
| 11 |
+
"mlp_ratio": 5.3334,
|
| 12 |
+
"pe_mode": "rope",
|
| 13 |
+
"share_mod": true,
|
| 14 |
+
"initialization": "scaled",
|
| 15 |
+
"qk_rms_norm": true,
|
| 16 |
+
"qk_rms_norm_cross": true
|
| 17 |
+
}
|
| 18 |
+
}
|
ckpts/tex_slat_denoiser.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f1f54f144509f5660aadf2f9a766457a66e8c1b96f88c6d3d5ccfff1e548d41
|
| 3 |
+
size 5169278120
|
pipeline.json
ADDED
|
@@ -0,0 +1,131 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "Alchemy3DInstructPipeline",
|
| 3 |
+
"args": {
|
| 4 |
+
"models": {
|
| 5 |
+
"sparse_structure_encoder": "microsoft/TRELLIS-image-large/ckpts/ss_enc_conv3d_16l8_fp16",
|
| 6 |
+
"sparse_structure_decoder": "microsoft/TRELLIS-image-large/ckpts/ss_dec_conv3d_16l8_fp16",
|
| 7 |
+
"sparse_structure_flow_model": "ckpts/sparse_structure_denoiser",
|
| 8 |
+
"shape_slat_encoder": "microsoft/TRELLIS.2-4B/ckpts/shape_enc_next_dc_f16c32_fp16",
|
| 9 |
+
"shape_slat_decoder": "microsoft/TRELLIS.2-4B/ckpts/shape_dec_next_dc_f16c32_fp16",
|
| 10 |
+
"shape_slat_flow_model": "ckpts/shape_slat_denoiser",
|
| 11 |
+
"tex_slat_encoder": "microsoft/TRELLIS.2-4B/ckpts/tex_enc_next_dc_f16c32_fp16",
|
| 12 |
+
"tex_slat_decoder": "microsoft/TRELLIS.2-4B/ckpts/tex_dec_next_dc_f16c32_fp16",
|
| 13 |
+
"tex_slat_flow_model": "ckpts/tex_slat_denoiser"
|
| 14 |
+
},
|
| 15 |
+
"sparse_structure_sampler": {
|
| 16 |
+
"name": "FlowEulerEditGuidanceIntervalSampler",
|
| 17 |
+
"args": {
|
| 18 |
+
"sigma_min": 1e-5
|
| 19 |
+
},
|
| 20 |
+
"params": {
|
| 21 |
+
"steps": 12,
|
| 22 |
+
"guidance_strength": 7.5,
|
| 23 |
+
"guidance_rescale": 0.7,
|
| 24 |
+
"guidance_interval": [0.6, 1.0],
|
| 25 |
+
"rescale_t": 5.0
|
| 26 |
+
}
|
| 27 |
+
},
|
| 28 |
+
"shape_slat_sampler": {
|
| 29 |
+
"name": "FlowEulerSparseEditGuidanceIntervalSampler",
|
| 30 |
+
"args": {
|
| 31 |
+
"sigma_min": 1e-5
|
| 32 |
+
},
|
| 33 |
+
"params": {
|
| 34 |
+
"steps": 12,
|
| 35 |
+
"guidance_strength": 7.5,
|
| 36 |
+
"guidance_rescale": 0.5,
|
| 37 |
+
"guidance_interval": [0.6, 1.0],
|
| 38 |
+
"rescale_t": 3.0
|
| 39 |
+
}
|
| 40 |
+
},
|
| 41 |
+
"tex_slat_sampler": {
|
| 42 |
+
"name": "FlowEulerSparseEditGuidanceIntervalSampler",
|
| 43 |
+
"args": {
|
| 44 |
+
"sigma_min": 1e-5
|
| 45 |
+
},
|
| 46 |
+
"params": {
|
| 47 |
+
"steps": 12,
|
| 48 |
+
"guidance_strength": 1.0,
|
| 49 |
+
"guidance_rescale": 0.0,
|
| 50 |
+
"guidance_interval": [0.6, 0.9],
|
| 51 |
+
"rescale_t": 3.0
|
| 52 |
+
}
|
| 53 |
+
},
|
| 54 |
+
"shape_slat_normalization": {
|
| 55 |
+
"mean": [
|
| 56 |
+
0.781296, 0.018091, -0.495192, -0.558457, 1.060530, 0.093252, 1.518149, -0.933218,
|
| 57 |
+
-0.732996, 2.604095, -0.118341, -2.143904, 0.495076, -2.179512, -2.130751, -0.996944,
|
| 58 |
+
0.261421, -2.217463, 1.260067, -0.150213, 3.790713, 1.481266, -1.046058, -1.523667,
|
| 59 |
+
-0.059621, 2.220780, 1.621212, 0.877230, 0.567247, -3.175944, -3.186688, 1.578665
|
| 60 |
+
],
|
| 61 |
+
"std": [
|
| 62 |
+
5.972266, 4.706852, 5.445010, 5.209927, 5.320220, 4.547237, 5.020802, 5.444004,
|
| 63 |
+
5.226681, 5.683095, 4.831436, 5.286469, 5.652043, 5.367606, 5.525084, 4.730578,
|
| 64 |
+
4.805265, 5.124013, 5.530808, 5.619001, 5.103930, 5.417670, 5.269677, 5.547194,
|
| 65 |
+
5.634698, 5.235274, 6.110351, 5.511298, 6.237273, 4.879207, 5.347008, 5.405691
|
| 66 |
+
]
|
| 67 |
+
},
|
| 68 |
+
"tex_slat_normalization": {
|
| 69 |
+
"mean": [
|
| 70 |
+
3.501659, 2.212398, 2.226094, 0.251093, -0.026248, -0.687364, 0.439898, -0.928075,
|
| 71 |
+
0.029398, -0.339596, -0.869527, 1.038479, -0.972385, 0.126042, -1.129303, 0.455149,
|
| 72 |
+
-1.209521, 2.069067, 0.544735, 2.569128, -0.323407, 2.293000, -1.925608, -1.217717,
|
| 73 |
+
1.213905, 0.971588, -0.023631, 0.106750, 2.021786, 0.250524, -0.662387, -0.768862
|
| 74 |
+
],
|
| 75 |
+
"std": [
|
| 76 |
+
2.665652, 2.743913, 2.765121, 2.595319, 3.037293, 2.291316, 2.144656, 2.911822,
|
| 77 |
+
2.969419, 2.501689, 2.154811, 3.163343, 2.621215, 2.381943, 3.186697, 3.021588,
|
| 78 |
+
2.295916, 3.234985, 3.233086, 2.260140, 2.874801, 2.810596, 3.292720, 2.674999,
|
| 79 |
+
2.680878, 2.372054, 2.451546, 2.353556, 2.995195, 2.379849, 2.786195, 2.775190
|
| 80 |
+
]
|
| 81 |
+
},
|
| 82 |
+
"text_cond_model": {
|
| 83 |
+
"name": "TextEncoder",
|
| 84 |
+
"args": {
|
| 85 |
+
"text_model": "Qwen/Qwen3.5-2B",
|
| 86 |
+
"mode": "ss_editing"
|
| 87 |
+
}
|
| 88 |
+
},
|
| 89 |
+
"ss_text_cond": {
|
| 90 |
+
"mode": "ss_editing",
|
| 91 |
+
"cond_projection": {
|
| 92 |
+
"model": {
|
| 93 |
+
"name": "TextProjection",
|
| 94 |
+
"args": {
|
| 95 |
+
"txt_dim": 2048,
|
| 96 |
+
"embedding_dim": 1024
|
| 97 |
+
}
|
| 98 |
+
},
|
| 99 |
+
"checkpoint": "ckpts/sparse_structure_cond_projection.safetensors"
|
| 100 |
+
}
|
| 101 |
+
},
|
| 102 |
+
"shape_text_cond": {
|
| 103 |
+
"mode": "shape_editing",
|
| 104 |
+
"cond_projection": {
|
| 105 |
+
"model": {
|
| 106 |
+
"name": "TextProjection",
|
| 107 |
+
"args": {
|
| 108 |
+
"txt_dim": 2048,
|
| 109 |
+
"embedding_dim": 1024
|
| 110 |
+
}
|
| 111 |
+
},
|
| 112 |
+
"checkpoint": "ckpts/shape_slat_cond_projection.safetensors"
|
| 113 |
+
}
|
| 114 |
+
},
|
| 115 |
+
"tex_text_cond": {
|
| 116 |
+
"mode": "texture_editing",
|
| 117 |
+
"cond_projection": {
|
| 118 |
+
"model": {
|
| 119 |
+
"name": "TextProjection",
|
| 120 |
+
"args": {
|
| 121 |
+
"txt_dim": 2048,
|
| 122 |
+
"embedding_dim": 1024
|
| 123 |
+
}
|
| 124 |
+
},
|
| 125 |
+
"checkpoint": "ckpts/tex_slat_cond_projection.safetensors"
|
| 126 |
+
}
|
| 127 |
+
},
|
| 128 |
+
"shape_resolution": 512,
|
| 129 |
+
"ss_occupancy_resolution": 64
|
| 130 |
+
}
|
| 131 |
+
}
|