Overview
Hi, here is FastFold v0.2.0. Compared to the previous version, now we inference faster and with less memory and support multimer inference.
What's Changed
- Save up to 75% GPU memory, able to inference sequence containing more than 10000 residues in bf16.
- Better softmax and layernorm kernel optimization based on Triton, at least 25% faster than previous version.
- Faster data processing, about 3 times faster on monomer, about 3N times faster on multimer with N sequence.
- Support multimer inference.
Installation
- Build from source
- Conda (recommended)
- Pypi
- Docker
Have a nice trip!