google-deepmind / google-deepmind/alphafold
Random seed during inference
Open
- Dominant language
- Python
- Stars
- 14.9k
- Forks
- 2.9k
- PR merge metrics
- No merged PRs in 30d
Description
I thought random numbers are required to initialize a model during training. So, why does AlphaFold require a random seed during inference? Where is this randomness introduced? Are these models deterministic?
Contributor guide
Assessment
This issue has not been assessed yet.