
esm
Uses the Biohub esm Python SDK for ESM3 protein generation, ESMC embeddings
by k-dense-aiFree
Description
esm is an agent skill for Biohub's esm Python SDK, which covers ESM3 protein generation, ESMC sequence embeddings and ESMFold2 all-atom structure prediction. It applies to local model inference and to Biohub's hosted services. Use it for ESM3 masked multimodal generation, for ESMC sequence representations, and for ESMFold2 folding. The former Forge platform has moved to biohub.ai, and the SDK's class names still contain the word Forge. The skill targets the released esm 3.4.1.post1 package, which it verified against the wheel and the current official documentation. It also separates related packages that people often confuse. fair-esm is Meta's separate, older ESM2, ESMFold and ESM-IF distribution. Both packages import as esm, so the skill says to install them in separate environments. The legacy esm.pretrained interface is not the Biohub ESMC interface, and Hugging Face Transformers has its own native ESMC implementation with different version requirements and output types. The skill's setup section explains how to choose a model. It is part of the k-dense-ai scientific agent skills repository on GitHub, which has about 49,600 stars.…
Highlights
- ESM3 masked multimodal protein generation
- ESMC sequence embeddings
- ESMFold2 all-atom folding
- Local inference and Biohub hosted services
- Separates fair-esm and Transformers ESMC
Ratings & Reviews
No ratings yet
Be the first to rate esm.
People also visit

pytorch-patterns
PyTorch deep learning patterns and best practices for building robust
esm FAQ
What can the SDK do?
Generate protein sequences with ESM3, compute ESMC sequence embeddings, and predict all-atom structures with ESMFold2.
Where did Forge go?
The former Forge platform has moved to biohub.ai. Some SDK class names still contain Forge.
Is fair-esm the same package?
No. fair-esm is Meta's older ESM2, ESMFold and ESM-IF distribution. Both import as esm, so the skill says to keep them in separate environments.
Which version does it target?
The released esm 3.4.1.post1 package, verified against its wheel and the current documentation.
How do I install it?
Run npx skills add https://github.com/k-dense-ai/scientific-agent-skills --skill esm.
Is it checked for security?
Its security audits show passes from Gen Agent Trust Hub, Socket and Snyk.
What are the best esm alternatives?
Popular alternatives to esm include diffdock, gget and pytorch-patterns. See all esm alternatives