{"id":28169881,"url":"https://github.com/zavalit/genmoai-models","last_synced_at":"2026-07-29T07:31:27.449Z","repository":{"id":260766110,"uuid":"881982454","full_name":"zavalit/genmoai-models","owner":"zavalit","description":"genmoai models fork","archived":false,"fork":false,"pushed_at":"2024-11-02T12:12:03.000Z","size":7116,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2024-11-02T13:18:37.047Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/zavalit.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-11-01T16:19:49.000Z","updated_at":"2024-11-02T12:12:06.000Z","dependencies_parsed_at":"2024-11-02T13:28:47.982Z","dependency_job_id":null,"html_url":"https://github.com/zavalit/genmoai-models","commit_stats":null,"previous_names":["zavalit/genmoai-models"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zavalit%2Fgenmoai-models","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zavalit%2Fgenmoai-models/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zavalit%2Fgenmoai-models/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zavalit%2Fgenmoai-models/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/zavalit","download_url":"https://codeload.github.com/zavalit/genmoai-models/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":254374525,"owners_count":22060615,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-05-15T16:18:53.616Z","updated_at":"2025-10-13T15:41:01.975Z","avatar_url":"https://github.com/zavalit.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Mochi 1\n[Blog](https://www.genmo.ai/blog) | [Hugging Face](https://huggingface.co/genmo/mochi-1-preview) | [Playground](https://www.genmo.ai/play) | [Careers](https://jobs.ashbyhq.com/genmo)\n\nA state of the art video generation model by [Genmo](https://genmo.ai).\n\nhttps://github.com/user-attachments/assets/4d268d02-906d-4cb0-87cc-f467f1497108\n\n## Overview\n\nMochi 1 preview is an open state-of-the-art video generation model with high-fidelity motion and strong prompt adherence in preliminary evaluation. This model dramatically closes the gap between closed and open video generation systems. We’re releasing the model under a permissive Apache 2.0 license. Try this model for free on [our playground](https://genmo.ai/play).\n\n## Installation\n\nInstall using [uv](https://github.com/astral-sh/uv):\n\n```bash\ngit clone https://github.com/genmoai/models\ncd models \npip install uv\nuv venv .venv\nsource .venv/bin/activate\nuv pip install setuptools\nuv pip install -e . --no-build-isolation\n```\n\nIf you want to install flash attention, you can use:\n```\nuv pip install -e .[flash] --no-build-isolation\n```\n\nYou will also need to install [FFMPEG](https://www.ffmpeg.org/) to turn your outputs into videos.\n\n## Download Weights\n\nUse [download_weights.py](scripts/download_weights.py) to download the model + decoder to a local directory. Use it like this:\n```\npython3 ./scripts/download_weights.py \u003cpath_to_downloaded_directory\u003e\n```\n\nOr, directly download the weights from [Hugging Face](https://huggingface.co/genmo/mochi-1-preview/tree/main) or via `magnet:?xt=urn:btih:441da1af7a16bcaa4f556964f8028d7113d21cbb\u0026dn=weights\u0026tr=udp://tracker.opentrackr.org:1337/announce` to a folder on your computer.\n\n## Running\n\nStart the gradio UI with\n\n```bash\npython3 ./demos/gradio_ui.py --model_dir \"\u003cpath_to_downloaded_directory\u003e\"\n```\n\nOr generate videos directly from the CLI with\n\n```bash\npython3 ./demos/cli.py --model_dir \"\u003cpath_to_downloaded_directory\u003e\"\n```\n\nReplace `\u003cpath_to_downloaded_directory\u003e` with the path to your model directory.\n\n## API\n\nThis repository comes with a simple, composable API, so you can programmatically call the model. You can find a full example [here](demos/api_example.py). But, roughly, it looks like this:\n\n```python\nfrom genmo.mochi_preview.pipelines import (\n    DecoderModelFactory,\n    DitModelFactory,\n    MochiSingleGPUPipeline,\n    T5ModelFactory,\n    linear_quadratic_schedule,\n)\n\npipeline = MochiSingleGPUPipeline(\n    text_encoder_factory=T5ModelFactory(),\n    dit_factory=DitModelFactory(\n        model_path=f\"{MOCHI_DIR}/dit.safetensors\", model_dtype=\"bf16\"\n    ),\n    decoder_factory=DecoderModelFactory(\n        model_path=f\"{MOCHI_DIR}/vae.safetensors\",\n    ),\n    cpu_offload=True,\n    decode_type=\"tiled_full\",\n)\n\nvideo = pipeline(\n    height=480,\n    width=848,\n    num_frames=31,\n    num_inference_steps=64,\n    sigma_schedule=linear_quadratic_schedule(64, 0.025),\n    cfg_schedule=[4.5] * 64,\n    batch_cfg=False,\n    prompt=\"your favorite prompt here ...\",\n    negative_prompt=\"\",\n    seed=12345,\n)\n```\n\n## Model Architecture\n\nMochi 1 represents a significant advancement in open-source video generation, featuring a 10 billion parameter diffusion model built on our novel Asymmetric Diffusion Transformer (AsymmDiT) architecture. Trained entirely from scratch, it is the largest video generative model ever openly released. And best of all, it’s a simple, hackable architecture. Additionally, we are releasing an inference harness that includes an efficient context parallel implementation. \n\nAlongside Mochi, we are open-sourcing our video AsymmVAE. We use an asymmetric encoder-decoder structure to build an efficient high quality compression model. Our AsymmVAE causally compresses videos to a 128x smaller size, with an 8x8 spatial and a 6x temporal compression to a 12-channel latent space. \n\n### AsymmVAE Model Specs\n|Params \u003cbr\u003e Count | Enc Base \u003cbr\u003e  Channels | Dec Base \u003cbr\u003e Channels |Latent \u003cbr\u003e Dim | Spatial \u003cbr\u003e Compression | Temporal \u003cbr\u003e Compression | \n|:--:|:--:|:--:|:--:|:--:|:--:|\n|362M   | 64  | 128  | 12   | 8x8   | 6x   | \n\nAn AsymmDiT efficiently processes user prompts alongside compressed video tokens by streamlining text processing and focusing neural network capacity on visual reasoning. AsymmDiT jointly attends to text and visual tokens with multi-modal self-attention and learns separate MLP layers for each modality, similar to Stable Diffusion 3. However, our visual stream has nearly 4 times as many parameters as the text stream via a larger hidden dimension. To unify the modalities in self-attention, we use non-square QKV and output projection layers. This asymmetric design reduces inference memory requirements.\nMany modern diffusion models use multiple pretrained language models to represent user prompts. In contrast, Mochi 1 simply encodes prompts with a single T5-XXL language model.\n\n### AsymmDiT Model Specs\n|Params \u003cbr\u003e Count | Num \u003cbr\u003e Layers | Num \u003cbr\u003e Heads | Visual \u003cbr\u003e Dim | Text \u003cbr\u003e Dim | Visual \u003cbr\u003e Tokens | Text \u003cbr\u003e Tokens | \n|:--:|:--:|:--:|:--:|:--:|:--:|:--:|\n|10B   | 48   | 24   | 3072   | 1536   | 44520   |   256   |\n\n## Hardware Requirements\n\nThe model requires at least 4 H100 GPUs to run. We welcome contributions from the community to reduce this requirement.\n\n## Safety\nGenmo video models are general text-to-video diffusion models that inherently reflect the biases and preconceptions found in their training data. While steps have been taken to limit NSFW content, organizations should implement additional safety protocols and careful consideration before deploying these model weights in any commercial services or products.\n\n## Limitations\nUnder the research preview, Mochi 1 is a living and evolving checkpoint. There are a few known limitations. The initial release generates videos at 480p today. In some edge cases with extreme motion, minor warping and distortions can also occur. Mochi 1 is also optimized for photorealistic styles so does not perform well with animated content. We also anticipate that the community will fine-tune the model to suit various aesthetic preferences.\n\n## Related Work\n- [ComfyUI-MochiWrapper](https://github.com/kijai/ComfyUI-MochiWrapper) adds ComfyUI support for Mochi. The integration of Pytorch's SDPA attention was taken from their repository.\n\n\n## BibTeX\n```\n@misc{genmo2024mochi,\n      title={Mochi},\n      author={Genmo Team},\n      year={2024}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzavalit%2Fgenmoai-models","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fzavalit%2Fgenmoai-models","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzavalit%2Fgenmoai-models/lists"}