zai-org/CogVideo

Health v2 · maintenance only

Adoption not part of the score

12977 stars · 1333 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

CogVideo/CogVideoX is an open-source family of text-to-video and image-to-video generation models from Zhipu AI (THUDM), with inference and fine-tuning code in Python. It includes model weights on Hugging Face, diffusers integration, LoRA fine-tuning, and demo tooling.

Use cases

• generate videos from a text prompt
• convert an image into a short video
• fine-tune a video generation model on custom data
• run text-to-video inference on a local GPU
• compare open-source alternatives to Sora
• build an app that animates still images

When to choose

• you need open weights for text-to-video or image-to-video generation
• you want to fine-tune a video diffusion model with LoRA on consumer GPUs
• you want diffusers-compatible video generation models

When to avoid

• you need commercial-grade, hosted video generation without managing GPUs
• you only need image generation, not video
• you lack a capable GPU for inference or training

Facets

library · maturity active

machine-learning deep-learning video-processing image-processing llm-inference llm-training artificial-intelligence deep-learning image-processing machine-learning python text-to-video image-to-video video-generation diffusion-models cogvideox sora model-weights diffusers video gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="zai-org/CogVideo")