Sergey O
Sergey O
Hi @oliverdutton ! Really cool contribution. Mind we try add it to colabfold? We already have fused attention and bfloat16 integrated into monomer model. Will be interesting to try flash...
Old GPUs do not support Unified Memory. Try remove these lines from run script: ``` 'TF_FORCE_UNIFIED_MEMORY': '1', 'XLA_PYTHON_CLIENT_MEM_FRACTION': '4.0', ```
This appears to be an issue with the latest version of openmm. Where before it would exit once converged (or reached some maximum number of iterations), now it can keep...
Can you provide more details of what and how you are running thing? Looks like there is mention of colabfold in the error message...
Yes. Even within the same model, changing the random seed will drastically result in varying quality.
For alphafold-multimer putting them on a line, shouldn't be a problem, but for alphafold-ptm... this can be an issue. I'll see if I can fix this in the beta version!...
one of the errors was coming from the citations.py. I fixed it in the beta branch: https://github.com/sokrypton/ColabFold/commit/80aef123bc0bb7c73468d478a35693459ba3506e
Can you share the exact error you are getting. Also, are you using Google Chrome?
this has been merged into "beta"
Does anyone have a new updated clear_cache() function, the solutions posted with latest jax are not able to resolve memory leaks.... :(