Merge branch 'main' into Dao-AILab/main

26f4b5fb · Woosuk Kwon · 5018ac6a · 12375706 · 5018ac6a · 5018ac6a
Commit 26f4b5fb authored Jul 31, 2024 by Woosuk Kwon
15 changed files
--- a/flash_attn/ops/triton/k_activations.py
+++ b/flash_attn/ops/triton/k_activations.py
--- a/flash_attn/ops/triton/layer_norm.py
+++ b/flash_attn/ops/triton/layer_norm.py
--- a/flash_attn/ops/triton/linear.py
+++ b/flash_attn/ops/triton/linear.py
--- a/flash_attn/ops/triton/mlp.py
+++ b/flash_attn/ops/triton/mlp.py
--- a/flash_attn/ops/triton/rotary.py
+++ b/flash_attn/ops/triton/rotary.py
--- a/flash_attn/utils/__init__.py
+++ b/flash_attn/utils/__init__.py
--- a/flash_attn/utils/benchmark.py
+++ b/flash_attn/utils/benchmark.py
--- a/flash_attn/utils/distributed.py
+++ b/flash_attn/utils/distributed.py
--- a/flash_attn/utils/generation.py
+++ b/flash_attn/utils/generation.py
--- a/flash_attn/utils/pretrained.py
+++ b/flash_attn/utils/pretrained.py
--- a/setup.py
+++ b/setup.py
--- a/tests/test_flash_attn.py
+++ b/tests/test_flash_attn.py
--- a/flash_attn/__init__.py
+++ b/flash_attn/__init__.py
-__version__ = "2.6.3"
+__version__ = "2.6.0"
-from flash_attn.flash_attn_interface import (
+from vllm_flash_attn.flash_attn_interface import (
    flash_attn_func,
    flash_attn_kvpacked_func,
    flash_attn_qkvpacked_func,

--- a/flash_attn/flash_attn_interface.py
+++ b/flash_attn/flash_attn_interface.py
--- a/flash_attn/pyproject.toml
+++ b/flash_attn/pyproject.toml