ProGen2 Protein Language Model

ProGen2 autoregressive protein language model for protein sequence generation

License: ProGen2 is open source and free for academic and commercial use under a BSD-3-Clause license. Please refer to the license for full terms.

This generator is open source. Any third-party models, product names, or trademarks referenced are the property of their respective owners, and Proto is not affiliated with them.

---

Tools Used Tools Used Source Source Cite Cite

Go to Tool Page

proto-bio/proto-language/proto_language/generator/progen2_generator.py

View source

@article{nijkamp2023progen2,
  title={ProGen2: Exploring the boundaries of protein language models},
  author={Nijkamp, Erik and Ruffolo, Jeffrey A and Weinstein, Eli N and Naik, Nikhil and Madani, Ali},
  journal={Cell Systems},
  volume={14},
  number={11},
  pages={968--978},
  year={2023},
  publisher={Elsevier},
  doi={10.1016/j.cels.2023.10.002}
}

Copy citation

Protein sequence generator using ProGen2 autoregressive language model.

API Reference

ConfigProGen2GeneratorConfig Source

Configuration object for ProGen2Generator.

This class defines configuration parameters for the ProGen2 generator, which uses the ProGen2 protein language model to autoregressively generate protein sequences from prompt sequences.

Models are loaded from HuggingFace: https://huggingface.co/hugohrban/

For detailed information on ProGen2, see:

• HuggingFace: https://huggingface.co/hugohrban/
• GitHub: https://github.com/hugohrban/ProGen2-finetuning
• Original GitHub: https://github.com/enijkamp/progen2
• Original paper: https://www.cell.com/cell-systems/fulltext/S2405-4712(23)00272-7

prompts · List[string] · required

Prompt sequences for protein sequence generation

model_checkpoint · enum · default: "progen2-large"

ProGen2 model variant to load (e.g. progen2-large).

Options: progen2-small, progen2-medium, progen2-base, progen2-oas, progen2-large, progen2-BFD90, progen2-xlarge

local_path · string

Path to local model weights

device · string · default: "cuda"

GPU device to run ProGen2 on (e.g. ‘cuda’ or ‘cuda:0’).

temperature · number · default: "0.2"

Sharpness of sampling. Below 1 favors high-probability tokens; above 1 increases diversity.

top_p · number · default: "0.95"

Nucleus sampling cumulative probability cutoff. 1.0 disables nucleus sampling.

top_k · integer · default: "0"

At each step, restrict sampling to the k most probable tokens. Set to 0 to disable top-k truncation.

truncate_at_stop · boolean · default: "True"

Whether to truncate sequences at stop tokens

strip_special_tokens · boolean · default: "True"

Whether to strip start and stop tokens from final output

prepend_prompt · boolean · default: "True"

Whether to prepend prompt to generation

batch_size · integer · default: "1"

Number of sequences to process simultaneously on GPU

verbose · boolean · default: "False"

Whether to print verbose output

Usage

python

from proto_language.generator import ProGen2Generator, ProGen2GeneratorConfig
from proto_language.core import Segment

config = ProGen2GeneratorConfig(
    # Configure parameters here
)

generator = ProGen2Generator(config)

segment = Segment(length=100, sequence_type="protein")
generator.assign(segment)
generator.sample()

Metadata

Property	Value
Key	progen2
Class	ProGen2Generator
Category	autoregressive
Input Type	prompt
Uses GPU	True
Supported Sequence Types	protein
Allows Empty Start	False