Chirag Jain 50896ec574 Make nvcc threads configurable via environment variable (#885) | пре 9 месеци | |
---|---|---|
.. | ||
README.md | пре 1 година | |
interface.cpp | пре 2 година | |
setup.py | пре 9 месеци | |
xentropy_kernel.cu | пре 2 година |
This CUDA extension implements optimized cross-entropy loss, adapted from Apex's Xentropy. We make it work for bfloat16 and support in-place backward to save memory.
It has only been tested on A100s.
cd csrc/xentropy && pip install .
As of 2023-09-15, this extension is no longer used in the FlashAttention repo. We've instead switched to a Triton-based implementation. See the CrossEntropyLoss module for more details.