FerrisMind commited on
Commit
f394917
·
verified ·
1 Parent(s): 055f3f3

Upload 6 files

Browse files
Files changed (6) hide show
  1. LICENSE +21 -0
  2. README.md +99 -0
  3. checksums.json +19 -0
  4. manifest.json +18 -0
  5. model.safetensors +3 -0
  6. resolved_config.json +40 -0
LICENSE ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MIT License
2
+
3
+ Copyright (c) 2024 Tripo AI & Stability AI
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
README.md ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ pipeline_tag: image-to-3d
4
+ base_model: stabilityai/TripoSR
5
+ tags:
6
+ - luxrt
7
+ - rust
8
+ - candle
9
+ - safetensors
10
+ - canonical-weights
11
+ - 3d
12
+ - image-to-3d
13
+ - mesh-generation
14
+ - triposr
15
+ ---
16
+
17
+ # LuxRT TripoSR Canonical Weights
18
+
19
+ This repository contains the canonicalized `safetensors` export of the TripoSR checkpoint used by the LuxRT `3d-rs` runtime.
20
+
21
+ The package converts the original TripoSR checkpoint/config pair into a stable runtime layout with explicit provenance and integrity metadata.
22
+
23
+ ## Model Details
24
+
25
+ - Model family: `triposr`
26
+ - Canonical file: `model.safetensors`
27
+ - Tensor count: `549`
28
+ - Primary runtime target: LuxRT `3d-rs`
29
+ - Typical output: mesh export (`.obj`)
30
+
31
+ ## Included Files
32
+
33
+ - `model.safetensors`: canonical model weights for runtime loading
34
+ - `resolved_config.json`: normalized JSON snapshot of the effective TripoSR config
35
+ - `manifest.json`: provenance metadata for the canonical package
36
+ - `checksums.json`: integrity data for the files in this package
37
+
38
+ ## Usage With LuxRT
39
+
40
+ This folder is designed to be uploaded as a standalone Hugging Face model repository and then restored into:
41
+
42
+ ```text
43
+ <repo-root>/3d/canonical-weights/triposr/
44
+ ```
45
+
46
+ The LuxRT CLI will validate and use the package. The example below assumes you run it from the `3d-rs/` directory:
47
+
48
+ ```powershell
49
+ cargo run -p lux3d-cli -- inspect --repo-root .. triposr
50
+ cargo run -p lux3d-cli -- run --repo-root .. triposr --source ../tp/3d/TripoSR/examples/horse.png --mc-resolution 256 --mc-threshold 25.0 --output ../3d/_generated/triposr.obj
51
+ ```
52
+
53
+ ## Provenance
54
+
55
+ - Upstream source model: [`stabilityai/TripoSR`](https://huggingface.co/stabilityai/TripoSR)
56
+ - Canonicalization flow: `3d-rs/tools/python_baseline/normalize_weights.py`
57
+ - Raw source files expected by LuxRT before normalization:
58
+ - `3d/models/stabilityai-TripoSR/model.ckpt`
59
+ - `3d/models/stabilityai-TripoSR/config.yaml`
60
+
61
+ The exact source checksum set is recorded in `manifest.json`. The integrity of the uploaded package is recorded in `checksums.json`.
62
+
63
+ ## Intended Use
64
+
65
+ This package is intended for:
66
+
67
+ - single-image 3D reconstruction in LuxRT
68
+ - mesh generation workflows using TripoSR canonical weights
69
+ - reproducible local/runtime validation
70
+
71
+ This package is not meant to replace the full upstream TripoSR repository. It is the runtime-focused canonical weight package for LuxRT.
72
+
73
+ ## Limitations
74
+
75
+ - Verified LuxRT inference currently assumes CUDA.
76
+ - This package includes canonical runtime artifacts only.
77
+ - Texture baking and other extended upstream utilities are outside the scope of this package.
78
+
79
+ ## License
80
+
81
+ The upstream TripoSR model and code are distributed under MIT.
82
+
83
+ - Upstream model page: [`stabilityai/TripoSR`](https://huggingface.co/stabilityai/TripoSR)
84
+ - Local package notice: [LICENSE](LICENSE)
85
+
86
+ Retain upstream attribution when redistributing the package.
87
+
88
+ ## Citation
89
+
90
+ If you use TripoSR in research, cite the upstream work:
91
+
92
+ ```bibtex
93
+ @article{tripoSR2024,
94
+ title={TripoSR: Fast 3D Object Reconstruction from a Single Image},
95
+ author={Tochilkin, Dmitry and Pankratz, David and Liu, Zexiang and Huang, Zixuan and Letts, Adam and Li, Yangguang and Liang, Ding and Laforte, Christian and Jampani, Varun and Cao, Yan-Pei},
96
+ journal={arXiv preprint arXiv:2403.02151},
97
+ year={2024}
98
+ }
99
+ ```
checksums.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "files": [
3
+ {
4
+ "relative_path": "3d/canonical-weights/triposr/model.safetensors",
5
+ "sha256": "f72bb520b8b1a5639600ac818496f22d6ccb3b42d3942412bd1e2375ef780a2b",
6
+ "size_bytes": 1677170936
7
+ },
8
+ {
9
+ "relative_path": "3d/canonical-weights/triposr/resolved_config.json",
10
+ "sha256": "480a3f742ac7f8f89e77e6a96eb8d0cb8564031d0c0f6fd225f9da9ad74a6686",
11
+ "size_bytes": 1164
12
+ },
13
+ {
14
+ "relative_path": "3d/canonical-weights/triposr/manifest.json",
15
+ "sha256": "33a9e558fcff782df50c60d66eb43bc29815fb5d6175b5757df76fa190774e34",
16
+ "size_bytes": 667
17
+ }
18
+ ]
19
+ }
manifest.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "canonical_file": "3d/canonical-weights/triposr/model.safetensors",
3
+ "dtype_histogram": {
4
+ "float32": 549
5
+ },
6
+ "family": "triposr",
7
+ "normalizer_version": 1,
8
+ "raw_files": [
9
+ "3d/models/stabilityai-TripoSR/model.ckpt",
10
+ "3d/models/stabilityai-TripoSR/config.yaml"
11
+ ],
12
+ "resolved_config_file": "3d/canonical-weights/triposr/resolved_config.json",
13
+ "source_checksums": {
14
+ "3d/models/stabilityai-TripoSR/config.yaml": "74ca708ce086bf68e97709ea6b3d91f14717921c04691e84043f0eb8fcc68e62",
15
+ "3d/models/stabilityai-TripoSR/model.ckpt": "429e2c6b22a0923967459de24d67f05962b235f79cde6b032aa7ed2ffcd970ee"
16
+ },
17
+ "tensor_count": 549
18
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f72bb520b8b1a5639600ac818496f22d6ccb3b42d3942412bd1e2375ef780a2b
3
+ size 1677170936
resolved_config.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backbone": {
3
+ "attention_head_dim": 64,
4
+ "cross_attention_dim": 768,
5
+ "in_channels": 1024,
6
+ "num_attention_heads": 16,
7
+ "num_layers": 16
8
+ },
9
+ "backbone_cls": "tsr.models.transformer.transformer_1d.Transformer1D",
10
+ "cond_image_size": 512,
11
+ "decoder": {
12
+ "activation": "silu",
13
+ "in_channels": 120,
14
+ "n_hidden_layers": 9,
15
+ "n_neurons": 64
16
+ },
17
+ "decoder_cls": "tsr.models.network_utils.NeRFMLP",
18
+ "image_tokenizer": {
19
+ "pretrained_model_name_or_path": "facebook/dino-vitb16"
20
+ },
21
+ "image_tokenizer_cls": "tsr.models.tokenizers.image.DINOSingleImageTokenizer",
22
+ "post_processor": {
23
+ "in_channels": 1024,
24
+ "out_channels": 40
25
+ },
26
+ "post_processor_cls": "tsr.models.network_utils.TriplaneUpsampleNetwork",
27
+ "renderer": {
28
+ "density_activation": "exp",
29
+ "density_bias": -1.0,
30
+ "feature_reduction": "concat",
31
+ "num_samples_per_ray": 128,
32
+ "radius": 0.87
33
+ },
34
+ "renderer_cls": "tsr.models.nerf_renderer.TriplaneNeRFRenderer",
35
+ "tokenizer": {
36
+ "num_channels": 1024,
37
+ "plane_size": 32
38
+ },
39
+ "tokenizer_cls": "tsr.models.tokenizers.triplane.Triplane1DTokenizer"
40
+ }