rewire.it
model · family

ESM-2

ESM-2 is a family of protein sequence transformers that produce residue-level and sequence-level representations.

0 evaluations · 0 metric rows

Shared profile: ESM-2. This page retains the exact record and its evaluation context.

At a glance

Explanatory profile: source reviewed · Automated source review, 2026-09-16. This does not change the review status of its results.

Inputs, outputs and configuration
PropertyDescription and evidence
Training resourceUniRef-derived protein sequencesfacebookresearch/esm official source · README.md: Main models, Getting started, Compute embeddings and Pre-trained Models
Configuration distinctionThe catalogue 8M entry is not the 650M or 15B checkpointfacebookresearch/esm official source · README.md: Main models, Getting started, Compute embeddings and Pre-trained Models
Configuration in this record8Mfacebookresearch/esm official source · README.md: Main models, Getting started, Compute embeddings and Pre-trained Models
Model typeNot extracted or verified for this record.
Context limitsNot extracted or verified for this record.
AccessNot extracted or verified for this record.
Code licenceNot extracted or verified for this record.
Weights licenceNot extracted or verified for this record.

How it works

Conceptual procedure

Schematic of the documented input, computation and output; not an executable configuration.

Conceptual procedureProtein sequence. Then: Amino-acid tokens. Then: ESM-2 transformer. Then: Residue embeddings. Then: Pooling or task predictorProtein sequenceAmino-acid tokensESM-2 transformerResidue embeddingsPooling or task predictor
Read the diagram as text
  1. Protein sequence
  2. Amino-acid tokens
  3. ESM-2 transformer
  4. Residue embeddings
  5. Pooling or task predictor
facebookresearch/esm official source · README.md: Main models, Getting started, Compute embeddings and Pre-trained Models

Amino-acid tokens pass through a pretrained transformer. Hidden states can be retained for each residue or pooled for a whole protein; downstream tasks need an explicit scoring rule or predictor. ESMFold adds a structure-prediction system and is a separate pipeline.

facebookresearch/esm official source · README.md: Main models, Getting started, Compute embeddings and Pre-trained Models

Benchmarks and results

Release 2026-09-16-d74d282221a9 · 0 evaluations · 0 metric rows. Different protocols are not a single leaderboard.

No evaluations linked in this release.

Strengths and limitations

Strengths supported by sources

  • Embeddings can be extracted directly from individual sequences; the repository provides several model sizes.facebookresearch/esm official source · README.md: Main models, Getting started, Compute embeddings and Pre-trained Models

Limitations and conditions

  • Different parameter sizes, pooling methods and supervised heads are not interchangeable evaluations.facebookresearch/esm official source · README.md: Main models, Getting started, Compute embeddings and Pre-trained Models
Profile review details

Primary project documentation or paper inspected for the explanatory claims and cited locations. Reviewed coverage concerns this narrative, not complete metadata, independent reproduction or a performance ranking.

Stable record: catalog-model-esm-2

Sources and history

Release 2026-09-16-d74d282221a9 · Record review: discovered

Download this release
Technical metadata and extraction receipts

Stable ID: catalog-model-esm-2

areas
proteins-complexes
method types
foundation model
entity level
family
version
8M
reported name
ESM-2
access
Public checkpoint; small 8M variant suits a local pilot.
method type
foundation model
missing metadata
checkpoint revision: not_yet_extracted; training data: not_yet_extracted; licence: not_yet_extracted
Related records

Suggest a correction