Skip to content

GitLab

Commit 04fd6d34 authored by Gustaf Ahdritz's avatar Gustaf Ahdritz
Browse files

Add note about FlashAttention

parent 4b3a102f
Hide whitespace changes
Inline Side-by-side
Showing with 2 additions and 2 deletions
README.md
View file @ 04fd6d34
...@@ -176,8 +176,8 @@ For large-scale batch inference, we offer an optional tracing mode, which ...@@ -176,8 +176,8 @@ For large-scale batch inference, we offer an optional tracing mode, which
massively improves runtimes at the cost of a lengthy model compilation process. massively improves runtimes at the cost of a lengthy model compilation process.
To enable it, add `--trace_model` to the inference command. To enable it, add `--trace_model` to the inference command.
To get a speedup during inference, [FlashAttention](https://github.com/HazyResearch/flash-attention) To get a speedup during inference, enable [FlashAttention](https://github.com/HazyResearch/flash-attention)
enable FlashAttention in the config. in the config. Note that it appears to work best for sequences with < 1000 residues.
Input FASTA files containing multiple sequences are treated as complexes. In Input FASTA files containing multiple sequences are treated as complexes. In
this case, the inference script runs AlphaFold-Gap, a hack proposed this case, the inference script runs AlphaFold-Gap, a hack proposed
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment