GitLab

Bugfix: Corrected paths for just-in-time embedding generation

Fix for a bug in data_transforms which wouldn't allow creation of MSA mask if there is only input sequence in MSA.
- Set `max_msa_clusters=1` in model presets for allowing the input sequence to be a MSA cluster centre.

- In `seqemb_mode`, `process_pdb` loads sequence embedding for the PDB's protein, and a dummy MSA

- In `seq_emb` mode, the AlignmentRunner works only on generating templates.

Added the seq_emb features to the list of features to be processed by feature pipeline , if using seq_emb mode
- In `seq_emb` mode, add list of `seq_emb` features to `feature_names`

- `_process_seqemb_features` now returns a dictionary instead of a tensor.

- Bugfix: `torch` throws warnings when copying a tensor via initialization
- Added lambda to `.clone()` those tensors instead

- Turn on `seqemb` mode in `data`, `model`, and `globals` config when using `seqemb` training preset.
- Set configuration options specific for finetuning in general.

- Turn on `seqemb` mode in `data`, `model`, and `globals` config when using `seqemb` training preset.

Added passing of sequence embedding mode flag from `data_modules` to `data_pipeline` for training and inference pipelines.
- Passing the config.data.seqemb_mode.enabled flag to the FASTA, PDB, and MMCIF data pipelines.

- Use sequence embedding files when in `seqemb` mode.
- Make dummy MSA features for MMCIF when using `seqemb` mode.

- Added flag `no_column_attention` in evoformer config.
- Added check in `evoformer.py` to switch off `MSAColumnAttention` when the config flag `no_column_attention` is `True`.