Name		Name	Last commit message	Last commit date
parent directory ..
configs		configs
testdata		testdata
README.md		README.md
__init__.py		__init__.py
data.py		data.py
data_test.py		data_test.py
domains.py		domains.py
embed.py		embed.py
embed_test.py		embed_test.py
evaluation.py		evaluation.py
evaluation_test.py		evaluation_test.py
fine_tune.py		fine_tune.py
fine_tune_test.py		fine_tune_test.py
generate_data.py		generate_data.py
logging.py		logging.py
models.py		models.py
models_test.py		models_test.py
modules.py		modules.py
modules_test.py		modules_test.py
requirements.txt		requirements.txt
run.sh		run.sh
sampling.py		sampling.py
sampling_test.py		sampling_test.py
seq_utils.py		seq_utils.py
train.py		train.py
utils.py		utils.py
utils_test.py		utils_test.py

README.md

Code for building and leveraging generative models on unlabeled protein data.

Setup a virtual environment and run training for a few steps with run.sh:

sh protein_lm/run.sh

Implements BERT and autoregressive models for proteins, as described in Biological Structure and Function Emerge from Scaling Unsupervised Learning to 250 Million Protein Sequences and ProGen: Language Modeling for Protein Generation.