chore: bump torch to 2.4.0

.github/workflows/publish.yml

@@ -49,7 +49,7 @@ jobs:
       matrix:
           os: ['ubuntu-20.04']
           python-version: ['3.8', '3.9', '3.10', '3.11']
-          pytorch-version: ['2.3.1']  # Must be the most recent version that meets requirements-cuda.txt.
+          pytorch-version: ['2.4.0']  # Must be the most recent version that meets requirements-cuda.txt.
           cuda-version: ['11.8', '12.1']
 
     steps:

CMakeLists.txt

@@ -32,7 +32,7 @@ set(HIP_SUPPORTED_ARCHS "gfx906;gfx908;gfx90a;gfx940;gfx941;gfx942;gfx1030;gfx11
 # requirements.txt files and should be kept consistent.  The ROCm torch
 # versions are derived from Dockerfile.rocm
 #
-set(TORCH_SUPPORTED_VERSION_CUDA "2.3.1")
+set(TORCH_SUPPORTED_VERSION_CUDA "2.4.0")
 set(TORCH_SUPPORTED_VERSION_ROCM "2.5.0")
 
 #

Dockerfile

@@ -116,7 +116,7 @@ RUN --mount=type=bind,from=build,src=/workspace/dist,target=/aphrodite-workspace
     python3 -m pip install dist/*.whl --verbose
 
 RUN --mount=type=cache,target=/root/.cache/pip \
-    python3 -m pip install https://github.com/flashinfer-ai/flashinfer/releases/download/v0.0.9/flashinfer-0.0.9+cu121torch2.3-cp310-cp310-linux_x86_64.whl
+    python3 -m pip install https://github.com/flashinfer-ai/flashinfer/releases/download/v0.1.2/flashinfer-0.1.2+cu121torch2.4-cp310-cp310-linux_x86_64.whl
 #################### Aphrodite installation IMAGE ####################
 
 

environment.yaml

@@ -7,7 +7,7 @@ channels:
   - defaults
 dependencies:
   - python=3.11.*
-  - pytorch=2.3.1
+  - pytorch=2.4.0
   - pytorch-cuda=12.1.*
   - cuda-nvcc=12.1.*
   - cuda-libraries-dev=12.1.*

pyproject.toml

@@ -4,7 +4,7 @@ requires = [
     "ninja",
     "packaging",
     "setuptools >= 49.4.0",
-    "torch == 2.3.1",
+    "torch == 2.4.0",
     "wheel",
 ]
 build-backend = "setuptools.build_meta"

requirements-build.txt

@@ -3,5 +3,5 @@ cmake>=3.21
 ninja
 packaging
 setuptools>=49.4.0
-torch==2.3.1
+torch==2.4.0
 wheel

requirements-cuda.txt

@@ -3,8 +3,8 @@
 
 # Dependencies for NVIDIA GPUs
 nvidia-ml-py == 12.555.43
-torch == 2.3.1
-torchvision == 0.18.1  # for phi3v
-xformers == 0.0.27  # Requires torch 2.3.1
+torch == 2.4.0
+torchvision == 0.19  # for phi3v
+xformers == 0.0.27.post2  # Requires torch 2.4.0
 triton >= 2.2.1
-vllm-flash-attn == 2.5.9.post1 # Requires PyTorch 2.3.1
+vllm-flash-attn == 2.6.0 # Requires PyTorch 2.4.0