From 9cbbec558b3d156c88218e4067f8e3f5fd8abe2d Mon Sep 17 00:00:00 2001 From: Marius Brehler Date: Wed, 13 Nov 2024 14:21:25 +0000 Subject: [PATCH 1/2] [sharktank] Unpin NumPy --- .github/workflows/ci-llama-large-tests.yaml | 3 +-- .github/workflows/ci-llama-quick-tests.yaml | 3 +-- .github/workflows/ci-shark-platform.yml | 3 +-- .github/workflows/ci_eval.yaml | 3 +-- sharktank/requirements.txt | 2 +- 5 files changed, 5 insertions(+), 9 deletions(-) diff --git a/.github/workflows/ci-llama-large-tests.yaml b/.github/workflows/ci-llama-large-tests.yaml index d79031b8c..5645efd8a 100644 --- a/.github/workflows/ci-llama-large-tests.yaml +++ b/.github/workflows/ci-llama-large-tests.yaml @@ -71,8 +71,7 @@ jobs: # Test with pinned nightly releases, not what iree-turbine uses. pip install -f https://iree.dev/pip-release-links.html --upgrade \ iree-base-compiler==2.9.0rc20241108 \ - iree-base-runtime==2.9.0rc20241108 \ - "numpy<2.0" + iree-base-runtime==2.9.0rc20241108 - name: Run llama tests run: pytest sharktank/tests/models/llama/benchmark_amdgpu_test.py -v -s --run-all-llama --iree-hip-target=gfx942 --html=out/index.html diff --git a/.github/workflows/ci-llama-quick-tests.yaml b/.github/workflows/ci-llama-quick-tests.yaml index decd0aa96..585a759ac 100644 --- a/.github/workflows/ci-llama-quick-tests.yaml +++ b/.github/workflows/ci-llama-quick-tests.yaml @@ -72,8 +72,7 @@ jobs: # Test with pinned nightly releases, not what iree-turbine uses. pip install -f https://iree.dev/pip-release-links.html --upgrade \ iree-base-compiler==2.9.0rc20241108 \ - iree-base-runtime==2.9.0rc20241108 \ - "numpy<2.0" + iree-base-runtime==2.9.0rc20241108 - name: Run llama 8b tests run: pytest sharktank/tests/models/llama/benchmark_amdgpu_test.py -v -s --iree-hip-target=gfx942 --run-8b-llama diff --git a/.github/workflows/ci-shark-platform.yml b/.github/workflows/ci-shark-platform.yml index 445e2e448..4a371a94c 100644 --- a/.github/workflows/ci-shark-platform.yml +++ b/.github/workflows/ci-shark-platform.yml @@ -69,8 +69,7 @@ jobs: # This should eventually stabilize. Do the best we can for now. pip install -f https://iree.dev/pip-release-links.html --upgrade \ iree-base-compiler \ - iree-base-runtime \ - "numpy<2.0" + iree-base-runtime - name: Run LLM Integration Tests run: pytest -v build_tools/integration_tests/llm --log-cli-level=INFO diff --git a/.github/workflows/ci_eval.yaml b/.github/workflows/ci_eval.yaml index 7afaeb1fe..b23d43699 100644 --- a/.github/workflows/ci_eval.yaml +++ b/.github/workflows/ci_eval.yaml @@ -71,8 +71,7 @@ jobs: # This should eventually stabilize. Do the best we can for now. pip install -f https://iree.dev/pip-release-links.html --upgrade \ iree-base-compiler \ - iree-base-runtime \ - "numpy<2.0" + iree-base-runtime - name: Run perplexity test with vmfb run: pytest -n 8 -v -s sharktank/tests/evaluate/perplexity_vmfb_test.py --longrun --iree-device='hip://7' --iree-hip-target=gfx942 --iree-hal-target-backends=rocm --llama3-8b-f16-model-path=/data/llama3.1/8b/llama8b_f16.irpa --llama3-8b-tokenizer-path=/data/llama3.1/8b/tokenizer_config.json diff --git a/sharktank/requirements.txt b/sharktank/requirements.txt index dd8f14fb6..150303718 100644 --- a/sharktank/requirements.txt +++ b/sharktank/requirements.txt @@ -2,7 +2,7 @@ iree-turbine # Runtime deps. gguf==0.6.0 -numpy==1.26.3 +numpy # Needed for newer gguf versions (TODO: remove when gguf package includes this) # sentencepiece>=0.1.98,<=0.2.0 From c974d5774843c97df273a882f99242ce6a95ec7c Mon Sep 17 00:00:00 2001 From: Marius Brehler Date: Thu, 14 Nov 2024 14:26:18 +0100 Subject: [PATCH 2/2] [sharktank] Only unpin NumPy on non-Windows --- sharktank/requirements.txt | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/sharktank/requirements.txt b/sharktank/requirements.txt index 150303718..19e48f825 100644 --- a/sharktank/requirements.txt +++ b/sharktank/requirements.txt @@ -2,7 +2,8 @@ iree-turbine # Runtime deps. gguf==0.6.0 -numpy +numpy==1.26.3; sys_platform == 'win32' +numpy; sys_platform != 'win32' # Needed for newer gguf versions (TODO: remove when gguf package includes this) # sentencepiece>=0.1.98,<=0.2.0