search

facebookresearch / esm

Evolutionary Scale Modeling (esm): Pretrained language models for proteins
MIT License
3.16k stars 627 forks source link

Using the same input for the inverse_folding model will produce different outputs #202

Closed lupengk closed 2 years ago

lupengk commented 2 years ago

Problem description When I run the script "examples/inverse_folding/score_log_likelihoods.py" repeatedly, it always produces different output, and the output is very different

Reproduction steps run the script "examples/inverse_folding/score_log_likelihoods.py" repeatedly

python examples/inverse_folding/score_log_likelihoods.py examples/inverse_folding/data/5YH2.pdb \
    examples/inverse_folding/data/5YH2_mutated_seqs.fasta --chain C \
    --outpath examples/inverse_folding/output/5YH2_mutated_seqs_scores.csv

The output varies greatly

tomsercu commented 2 years ago

Setting the model in eval mode should fix this issue. It's being addressed in #206 Please feel free to re-open if that didn't properly address the problem! Thanks

lupengk commented 2 years ago

Great, thanks!