Repository Issue Activity (beta)

#151 Request for correction: ProteinMPNN manifest draft in OpenSciFlow

#150 [K-Lean] Lean 4 sorry-free contracts for protein sequence design — SequenceIdentity, ProteinRMSD, ShannonEntropy

#142 What should "--num_examples_per_epoch" be set to?

#149 ProteinMPNN appears to modify fixed target sequence during peptide design workflow

#148 Problem: --chain_id_jsonl and --fixed_positions_jsonl not working as expected (both chains designed)

#122 Combine "specify residues to design" and "homooligomer"

#147 the difference between the conditional_probs_only_backbone mode and the unconditional_probs_only mode

#145 [Documentation] Simpler explanation?

#144 homooligomer and different length

#107 No use of GPU?

#141 Best Practices for Large-Scale ProteinMPNN → FastRelax Runs: Are Per-Backbone Designs Truly Independent?

#25 Abnormal results

#139 Output interpretation

#138 Security audit results

#123 Difference between `log_probs` and `probs` when set `save_probs`

#136 Possible bug in model initialization (`num_decoder_layers`)

#135 Do amino acid types specified in the input backbone PDB file affect the designed result?

#132 Using ProteinMPNN *without* oxygens?

#133 Fixed positions JSONL only loads last entry

#131 How does ProteinMPNN select the sampled sequences to report in the output file?

#130 How process a batch data?

#127 why generate an empty .jsonl file?

#129 Question about list.csv HASH (mentioned in training/README.md)

#128 temperature

#83 Training model