Multi-Modal Foundation · biohub
ESMC 600M (Dec 2024)
A 600-million-parameter protein language model checkpoint for protein engineering and variant-effect analysis.
What ESMC 600M (Dec 2024) is
ESMC 600M (Dec 2024) is a 600-million-parameter protein language model developed by Biohub for applications across biology, medicine, and genomics. It serves as a widely adopted checkpoint designed to represent protein sequences for computational evaluation.
The checkpoint underpins research in protein engineering and variant-effect workflows, assisting researchers in studying sequence variations and designing altered proteins. This specific release provides weights formatted for the GitHub-compatible esm package and is retained to maintain backwards compatibility with existing pipelines.
Key facts
- Parameter count
- 600 million
- Model type
- Protein language model
- Release date
- December 2024
- Format
- GitHub-compatible esm package format
What you can use it for
- Protein Engineering
- Guides the evaluation and design of engineered protein sequences by leveraging learned sequence representations.
- Variant-Effect Workflows
- Assists in assessing the impact of sequence mutations and variations within genomic and biological contexts.
- Legacy Pipeline Execution
- Enables existing workflows that rely specifically on the GitHub esm package format to reproduce results without code changes.
Run it here
We list this model but cannot run it yet: superseded by biohub/ESMC-600M, already indexed.
Everything on this page still applies. Running it needs a loader for this repository, not a bigger machine.
Strengths and limits
Good at
- Provides a widely adopted 600-million-parameter checkpoint for protein representation.
- Maintains backwards compatibility for pipelines relying on the original esm package format.
Watch out for
- Intended primarily for backwards compatibility rather than new pipeline development.
- Uses the GitHub esm package format rather than standard Hugging Face model weight formats.
Inputs and outputs
You provide
Sequence or text
A protein or nucleotide sequence, or free text — whichever this model was trained on.
You get back
Mean-pooled representation, one vector per input.
Licence
Check the licence before commercial use · mit
The mit licence carries terms worth reading before you depend on it.
Questions
Can I use ESMC 600M (Dec 2024) commercially?
Check the licence before commercial use. The mit licence carries terms worth reading before you depend on it.
What is ESMC 600M (Dec 2024)?
It is a December 2024 release of Biohub's 600-million-parameter protein language model, packaged in the GitHub-compatible esm format.
What scientific tasks does this model support?
It supports biological, medical, and genomic research, particularly variant-effect analysis and protein engineering.
Why is this specific checkpoint maintained?
It is maintained for backwards compatibility with software and workflows configured for the original esm package distribution.
Written from ESMC 600M (Dec 2024)'s own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.