try flash attn

Signed-off-by: Max Jeblick <maximilianjeblick@gmail.com>

Author: maxjeblick

.github/workflows/test.yml

@@ -16,14 +16,6 @@ jobs:
         with:
           python-version: 3.10.11
 
-      - name: Setup CUDA
-        uses: Jimver/cuda-toolkit@v0.2.16
-        with:
-          cuda: '12.5.0'
-
-      - name: Set CUDA_HOME
-        run: echo "CUDA_HOME=/usr/local/cuda" >> $GITHUB_ENV
-
       - name: Install uv
         uses: astral-sh/setup-uv@v6
         with:

Makefile

@@ -42,7 +42,7 @@ reports:
 .PHONY: test
 test: reports
 	$(UV) pip install optimum-quanto
-	$(UV) pip install flash-attn
+	$(UV) pip install flash-attn --no-build-isolation
 	PYTHONPATH=. \
 	$(UV) run pytest \
 		--cov-report xml:reports/coverage.xml \