bitsandbytes
k-bit quantization (QLoRA) and 8-bit optimizers for PyTorch
homepage ↗ pypi: bitsandbytes github: bitsandbytes-foundation/bitsandbytes
Available in
| Overlay | Newest | Ebuilds | Last activity | |
|---|---|---|---|---|
| stuff GitHub ↗ | 0.50.0 | 2 | 3 d | details › |
Versions & arches
Use flags of 0.50.0
- rocm Build the HIP/ROCm GPU backend (COMPUTE_BACKEND=hip) instead of the CPU backend