GitLab

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.

update train_openfold to handle val_mmcif_cache;update datamodule to handle if validation mmcif cache is not provided

Fixes for memory usage in multimer dataloader, spatial cropping, and tensor type conversions within model.

remove number of chains filter; no longer hard code is_multimer in multime classes; remove recycling dimesion in ground truth features

finished constructing multimer dataset object and now return both all_seq_features and ground truth structure

modified OpenFoldSingleDataset class so that it takes mmcif_cache information; created a new OpenFoldMultimerDataset class that will create suitable data input for multimer training