e1jax 0.1.0.dev0

An implementation of the E1 protein language model family with Jax/Equinox.

e1-jax

A minimal implementation of the E1 protein language model family with Jax/Equinox. Logits and embeddings match those given by authors repository.

Dependencies are managed with uv. To install them, run uv sync.

Currently, only single sequence inference is supported.

Example

The implementation is compatible with equinox.filter_{vmap, jit} for batched and jitted inference.

The model can be one of E1-150m, E1-300m, E1-600m.

import e1jax

seq = "AAAAA?C"
pad_length = 10
tokenized = e1jax.tokenize(seq)
tokenized = e1jax.pad_and_mask(tokenized, pad_length=pad_length)

model = e1jax.E1.from_pretrained("E1-300m")

logits, embeddings = model(**tokenized._asdict())

# to remove boundary tokens and padding
lb, rb = 2, -2-pad_length
logits, embeddings = logits[lb:rb], embeddings[lb:rb]

assert logits.shape[0] == len(seq)
assert embeddings.shape[0] == len(seq)

Citations

 @article{Jain_Beazer_Ruffolo_Bhatnagar_Madani_2025,
    title={E1: Retrieval-Augmented Protein Encoder Models},
    url={https://www.biorxiv.org/content/early/2025/11/13/2025.11.12.688125},
    DOI={10.1101/2025.11.12.688125},
    journal={bioRxiv},
    publisher={Cold Spring Harbor Laboratory},
    author={Jain, Sarthak and Beazer, Joel and Ruffolo, Jeffrey A and Bhatnagar, Aadyot and Madani, Ali},
    year={2025}
}