From 71f1a6876cfddf68e9676a3d6b6bfec31f282f40 Mon Sep 17 00:00:00 2001 From: JC1DA Date: Mon, 13 Jan 2025 13:51:05 -0800 Subject: [PATCH] Exclude llama-cpp-python 0.3.6 in testcases (#1096) Latest llama-cpp-python 0.3.6 breaks some of our testcases. Relevant issue: https://github.com/ggerganov/llama.cpp/issues/11197 (Different results returned using same prompt with temperature 0) Temporarily exclude this version until it's fixed Signed-off-by: Loc Huynh Co-authored-by: Loc Huynh --- .github/workflows/action_gpu_basic_tests.yml | 2 +- .github/workflows/action_plain_basic_tests.yml | 2 +- .github/workflows/ci_tests.yml | 4 ++-- .github/workflows/notebook_tests.yml | 2 +- 4 files changed, 5 insertions(+), 5 deletions(-) diff --git a/.github/workflows/action_gpu_basic_tests.yml b/.github/workflows/action_gpu_basic_tests.yml index 97190ec34..5b5a09bad 100644 --- a/.github/workflows/action_gpu_basic_tests.yml +++ b/.github/workflows/action_gpu_basic_tests.yml @@ -62,7 +62,7 @@ jobs: pip install accelerate echo "==============================" pip uninstall -y llama-cpp-python - CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75,!=0.2.84" + CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75,!=0.2.84,!=0.3.6" - name: Check GPU available run: | python -c "import torch; assert torch.cuda.is_available()" diff --git a/.github/workflows/action_plain_basic_tests.yml b/.github/workflows/action_plain_basic_tests.yml index 789a53ab9..0a2d0b6c3 100644 --- a/.github/workflows/action_plain_basic_tests.yml +++ b/.github/workflows/action_plain_basic_tests.yml @@ -40,7 +40,7 @@ jobs: pip install sentencepiece echo "==============================" pip uninstall -y llama-cpp-python - pip install "llama-cpp-python!=0.2.58,!=0.2.79,!=0.2.84" + pip install "llama-cpp-python!=0.2.58,!=0.2.79,!=0.2.84,!=0.3.6" echo "==============================" pip uninstall -y transformers pip install "transformers!=4.43.0,!=4.43.1,!=4.43.2,!=4.43.3" # Issue 965 diff --git a/.github/workflows/ci_tests.yml b/.github/workflows/ci_tests.yml index 1911f99fe..54453442d 100644 --- a/.github/workflows/ci_tests.yml +++ b/.github/workflows/ci_tests.yml @@ -57,7 +57,7 @@ jobs: - name: GPU pip installs run: | pip install accelerate - CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75,!=0.2.84" + CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75,!=0.2.84,!=0.3.6" - name: Check GPU available run: | python -c "import torch; assert torch.cuda.is_available()" @@ -153,7 +153,7 @@ jobs: echo "======================" nvcc --version echo "======================" - CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75" + CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75,!=0.3.6" - name: Check GPU available run: | python -c "import torch; assert torch.cuda.is_available()" diff --git a/.github/workflows/notebook_tests.yml b/.github/workflows/notebook_tests.yml index d2b19e622..fd217d321 100644 --- a/.github/workflows/notebook_tests.yml +++ b/.github/workflows/notebook_tests.yml @@ -60,7 +60,7 @@ jobs: - name: GPU pip installs run: | pip install accelerate - CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75,!=0.2.84" + CMAKE_ARGS="-DGGML_CUDA=on" pip install "llama-cpp-python!=0.2.58,!=0.2.75,!=0.2.84,!=0.3.6" - name: Check GPU available run: | python -c "import torch; assert torch.cuda.is_available()"