KaiyangZhou/CoOp
Prompt learning library for adapting vision-language foundation models like CLIP to downstream classification tasks.
Not currently ranked — collecting fresh signals.
star history
Provides implementations of CoOp and CoCoOp, two prompt learning methods for adapting large vision-language models (CLIP) to downstream datasets without full model fine-tuning. The techniques learn trainable prompt embeddings that combine with frozen pre-trained model features, enabling parameter-efficient transfer learning to new domains. Supports various downstream classification tasks including ImageNet and domain-shifted variants.
Frequently asked
- What is KaiyangZhou/CoOp?
- Prompt learning library for adapting vision-language foundation models like CLIP to downstream classification tasks.
- Is CoOp open source?
- Yes — KaiyangZhou/CoOp is open source, released under the MIT license.
- What language is CoOp written in?
- KaiyangZhou/CoOp is primarily written in Python.
- How popular is CoOp?
- KaiyangZhou/CoOp has 2.2k stars on GitHub.
- Where can I find CoOp?
- KaiyangZhou/CoOp is on GitHub at https://github.com/KaiyangZhou/CoOp.