Exclude llama-cpp-python 0.3.6 in testcases (#1096)

Latest llama-cpp-python 0.3.6 breaks some of our testcases.

Relevant issue: ggerganov/llama.cpp#11197
(Different results returned using same prompt with temperature 0)

Temporarily exclude this version until it's fixed

Signed-off-by: Loc Huynh <[email protected]>
Co-authored-by: Loc Huynh <[email protected]>

.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()"

.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

.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()"

.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()"