Cosmos
Diffusers
Safetensors
cosmos3_omni
nvidia
cosmos3
vllm
vllm-omni
sglang
sglang-diffusion
text, image, video, audio, and action generation
omnimodel
Instructions to use nvidia/Cosmos3-Super-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Cosmos
How to use nvidia/Cosmos3-Super-FP8 with Cosmos:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Diffusers
How to use nvidia/Cosmos3-Super-FP8 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("nvidia/Cosmos3-Super-FP8", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
Commit History
7 days ago
-
Fix PyTorch runtime link in model card (#9)
- Fix PyTorch runtime link in model card (d53fb82eb889b71905b812351e29cceea7b77ec2) Co-authored-by: Yao Xu <yaoxuhf@users.noreply.huggingface.co>3f62010
9 days ago
-
Regenerate Diffusers T2V example with Super FP8 (#8)
- Regenerate Diffusers T2V example with Super FP8 (3ab63bcfc601471c87fbddf8619a7ab5e48db65d) Co-authored-by: Yiming Zhao <ConstBob@users.noreply.huggingface.co>88d5a69 -
Sync Quality Benchmark from reviewed MR !13060
Verified signatureSync only the complete Quality Benchmark section from GitLab MR !13060 at eddc7e7a1f221b6f31ec6c107e32e91d8fc31bea. Preserve all other README sections and repository files unchanged.db533ab
10 days ago
-
Update sglang description; remove unused assets (#7)
- remove benchmark images; update sglang description (50e4ae6f1682c6244fd33b5137ea5ea32467348c) Co-authored-by: Mu-Ti Chung <mutic@users.noreply.huggingface.co>b4d3ff2 -
Integrate reviewed Cosmos3-Super-FP8 model card with current usage and performance
Verified signatureUse the reviewed MR overview, quality benchmarks, and training/evaluation datasets; retain current HF usage examples and performance measurements. Add the PNG relationship diagram, set release date to November 9, 2026, and correct precision and section links. Exclude internal-only sections. Existing shared subcards are unchanged.c20eece -
docs: fix Diffusers FP8 checkpoint restore (#4)
- docs: fix Diffusers FP8 checkpoint restore (9057e2e9a98ab6865803210ff6f01b825498e16d) Co-authored-by: Yiming Zhao <ConstBob@users.noreply.huggingface.co>382840d -
Update vLLM, SGLang and vLLM-Omni examples; refresh generated examples from FP8 (#6)
- Update vLLM, SGLang and vLLM-Omni examples; refresh generated examples from FP8 (5dae7f8589fc5e566bc7e0cfeb0f27c318213260) - Cap huggingface_hub below 2 and note the Rust toolchain requirement for the SGLang editable install (891647b053fd92ef727cad2593db4503cd26cfc2) - Drop the nonexistent [cli] extra from the huggingface_hub install lines (db9f24e2199fc3b7d2eb1e6762d36c7044e99f65) Co-authored-by: Wojciech Kutak <wkutak@users.noreply.huggingface.co>ef06c4d -
Add compact links to both Cosmos3 collections
Verified signature117a9c6
13 days ago
-
Fill missing GB300 Reasoner performance benchmarks (#5)
- Fill missing GB300 Reasoner performance benchmarks (02992315e9db8b30dabc87f4ad588bca0dc65006) - Remove obsolete unavailable-measurement note (f97cf3a55ec2638692868032938247a636a3c38d) Co-authored-by: Sergiy Fefilatyev <fsergiy@users.noreply.huggingface.co>c10c850
-
Add Hopper Generator inference benchmark results (#3)
- Add Hopper Generator inference benchmark results (9c6ede42f5abf2750b4c35f2cad099d0415cc770) - Add B300-based GB300 Action Policy estimate (c3ff73ab6da580fbfdeb304836e46d9b0b44e5ae) - Simplify GB300 estimate notation (e29bd924ec24c050ea6beaaeba4ddc1ab245ea8d) Co-authored-by: Sergiy Fefilatyev <fsergiy@users.noreply.huggingface.co>f5d7c44
-
Add Cosmos3-Super-FP8 inference performance benchmarks (#2)
- Add Cosmos3-Super-FP8 inference performance benchmarks (43cea8ec6c06aa87ce2fad09a0b796403c574cc2) Co-authored-by: Sergiy Fefilatyev <fsergiy@users.noreply.huggingface.co>4844b1c
-
Add Cosmos3 Super FP8 snapshot (#1)
- Track Super FP8 tokenizer and example assets (b0da4ea638c34b94c0f90236a195759736249bf4) - Add Cosmos3 Super FP8 snapshot files (8ee88c146703d02bad9dac91decbb5c4fa1b3552) Co-authored-by: Wojciech Kutak <wkutak@users.noreply.huggingface.co>4dd8d27
-
Fix model card image and video links to original model assets
Verified signature589722a -
Add FP8 model card based on nvidia/Cosmos3-Super
Verified signature628bfb1 -
initial commit
Verified signaturecd771d0