Genomic foundation models from Hugging Face.
Each model release lives in its own directory, with its own dependencies, training recipes, and evals. Run commands from inside that directory.
| Directory | Models | Contents |
|---|---|---|
carbon/ |
Carbon-500M, Carbon-3B, Carbon-8B | Zero-shot DNA evals, finetuning recipes, embedding analysis |
carbon-x/ |
Carbon-X | Midtraining, SFT and RL recipes, DNA understanding evals |
cd carbon
uv syncThe Carbon tech report is available on bioRxiv.
The Carbon-X blog post is available here.
Carbon is a joint collaboration between the research teams at Hugging Face, Zhongguancun Academy, and TIGEM/University of Naples “Federico II”.
Apache 2.0.