[Neuron][Kernel] NKI-based flash-attention kernel with paged KV cache (vllm-project#11277)

Signed-off-by: Liangfu Chen <[email protected]>
Co-authored-by: Jiangfei Duan <[email protected]>

Author: 2 people

.buildkite/run-neuron-test.sh

@@ -54,4 +54,4 @@ docker run --rm -it --device=/dev/neuron0 --device=/dev/neuron1 --network host \
        -e "NEURON_COMPILE_CACHE_URL=${NEURON_COMPILE_CACHE_MOUNT}" \
        --name "${container_name}" \
        ${image_name} \
-       /bin/bash -c "python3 /workspace/vllm/examples/offline_inference/neuron.py"
+       /bin/bash -c "python3 /workspace/vllm/examples/offline_inference/neuron.py && python3 -m pytest /workspace/vllm/tests/neuron/ -v --capture=tee-sys"