For the "Supervised variant prediction - training a classifier on the embeddings" part, I ran the command "python scripts/extract.py esm1v_t33_650M_UR90S_1 examples/data/P62593.fasta \
examples/data/P62593_emb_esm1v --repr_layers 33 --include mean" to get embeddings for the sample fasta file P62593.fasta and I compared my computed embeddings to the precomputed embeddings provided in https://dl.fbaipublicfiles.com/fair-esm/examples/P62593_reprs.tar.gz, I found that the embeddings for the same sequence are different. May I know what command you use to obtain the precomputed embeddings? Do you use the esm1v_t33_650M_UR90S_1?
I hope that you can help with my questions. Thank you in advance.
For the "Supervised variant prediction - training a classifier on the embeddings" part, I ran the command "python scripts/extract.py esm1v_t33_650M_UR90S_1 examples/data/P62593.fasta \ examples/data/P62593_emb_esm1v --repr_layers 33 --include mean" to get embeddings for the sample fasta file P62593.fasta and I compared my computed embeddings to the precomputed embeddings provided in https://dl.fbaipublicfiles.com/fair-esm/examples/P62593_reprs.tar.gz, I found that the embeddings for the same sequence are different. May I know what command you use to obtain the precomputed embeddings? Do you use the esm1v_t33_650M_UR90S_1?
I hope that you can help with my questions. Thank you in advance.