Move requirements into their own directory (#12547)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

.buildkite/nightly-benchmarks/scripts/run-nightly-benchmarks.sh

@@ -426,7 +426,7 @@ main() {
 
   pip install -U transformers
 
-  pip install -r requirements-dev.txt
+  pip install -r requirements/dev.txt
   which genai-perf
 
   # check storage

.buildkite/run-cpu-test.sh

@@ -35,7 +35,7 @@ function cpu_tests() {
   # Run basic model test
   docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
     set -e
-    pip install -r vllm/requirements-test.txt
+    pip install -r vllm/requirements/test.txt
     pytest -v -s tests/models/decoder_only/language -m cpu_model
     pytest -v -s tests/models/embedding/language -m cpu_model
     pytest -v -s tests/models/encoder_decoder/language -m cpu_model

.buildkite/test-pipeline.yaml

@@ -35,7 +35,7 @@ steps:
   fast_check: true
   no_gpu: True
   commands:
-  - pip install -r requirements-docs.txt
+  - pip install -r ../../requirements/docs.txt
   - SPHINXOPTS=\"-W\" make html
   # Check API reference (if it fails, you may have missing mock imports)
   - grep \"sig sig-object py\" build/html/api/inference_params.html

.github/workflows/publish.yml

@@ -50,7 +50,7 @@ jobs:
   #     matrix:
   #         os: ['ubuntu-20.04']
   #         python-version: ['3.9', '3.10', '3.11', '3.12']
-  #         pytorch-version: ['2.4.0']  # 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:

.github/workflows/scripts/build.sh

@@ -9,7 +9,7 @@ PATH=${cuda_home}/bin:$PATH
 LD_LIBRARY_PATH=${cuda_home}/lib64:$LD_LIBRARY_PATH
 
 # Install requirements
-$python_executable -m pip install -r requirements-build.txt -r requirements-cuda.txt
+$python_executable -m pip install -r requirements/build.txt -r requirements/cuda.txt
 
 # Limit the number of parallel jobs to avoid OOM
 export MAX_JOBS=1

.pre-commit-config.yaml

@@ -44,8 +44,8 @@ repos:
   rev: 0.6.2
   hooks:
     - id: pip-compile
-      args: [requirements-test.in, -o, requirements-test.txt]
+      args: [requirements/test.in, -o, requirements/test.txt]
-      files: ^requirements-test\.(in|txt)$
+      files: ^requirements/test\.(in|txt)$
 - repo: local
   hooks:
   - id: mypy-local

.readthedocs.yaml

@@ -18,4 +18,4 @@ formats: []
 # Optionally declare the Python requirements required to build your docs
 python:
   install:
-    - requirements: docs/requirements-docs.txt
+    - requirements: requirements/docs.txt

Dockerfile

@@ -55,10 +55,10 @@ RUN --mount=type=cache,target=/root/.cache/uv \
         uv pip install --index-url https://download.pytorch.org/whl/nightly/cu126 "torch==2.7.0.dev20250121+cu126" "torchvision==0.22.0.dev20250121";  \
     fi
 
-COPY requirements-common.txt requirements-common.txt
+COPY requirements/common.txt requirements/common.txt
-COPY requirements-cuda.txt requirements-cuda.txt
+COPY requirements/cuda.txt requirements/cuda.txt
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv pip install -r requirements-cuda.txt
+    uv pip install -r requirements/cuda.txt
 
 # cuda arch list used by torch
 # can be useful for both `dev` and `test`
@@ -76,14 +76,14 @@ FROM base AS build
 ARG TARGETPLATFORM
 
 # install build dependencies
-COPY requirements-build.txt requirements-build.txt
+COPY requirements/build.txt requirements/build.txt
 
 # This timeout (in seconds) is necessary when installing some dependencies via uv since it's likely to time out
 # Reference: https://github.com/astral-sh/uv/pull/1694
 ENV UV_HTTP_TIMEOUT=500
 
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv pip install -r requirements-build.txt
+    uv pip install -r requirements/build.txt
 
 COPY . .
 ARG GIT_REPO_CHECK=0
@@ -151,11 +151,11 @@ FROM base as dev
 # Reference: https://github.com/astral-sh/uv/pull/1694
 ENV UV_HTTP_TIMEOUT=500
 
-COPY requirements-lint.txt requirements-lint.txt
+COPY requirements/lint.txt requirements/lint.txt
-COPY requirements-test.txt requirements-test.txt
+COPY requirements/test.txt requirements/test.txt
-COPY requirements-dev.txt requirements-dev.txt
+COPY requirements/dev.txt requirements/dev.txt
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv pip install -r requirements-dev.txt
+    uv pip install -r requirements/dev.txt
 #################### DEV IMAGE ####################
 
 #################### vLLM installation IMAGE ####################
@@ -230,9 +230,9 @@ COPY examples examples
 # some issues w.r.t. JIT compilation. Therefore we need to
 # install build dependencies for JIT compilation.
 # TODO: Remove this once FlashInfer AOT wheel is fixed
-COPY requirements-build.txt requirements-build.txt
+COPY requirements/build.txt requirements/build.txt
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv pip install -r requirements-build.txt
+    uv pip install -r requirements/build.txt
 
 #################### vLLM installation IMAGE ####################
 
@@ -249,7 +249,7 @@ ENV UV_HTTP_TIMEOUT=500
 
 # install development dependencies (for testing)
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv pip install -r requirements-dev.txt
+    uv pip install -r requirements/dev.txt
 
 # install development dependencies (for testing)
 RUN --mount=type=cache,target=/root/.cache/uv \

Dockerfile.arm

@@ -26,18 +26,18 @@ WORKDIR /workspace
 ARG PIP_EXTRA_INDEX_URL="https://download.pytorch.org/whl/cpu"
 ENV PIP_EXTRA_INDEX_URL=${PIP_EXTRA_INDEX_URL}
 RUN --mount=type=cache,target=/root/.cache/pip \
-    --mount=type=bind,src=requirements-build.txt,target=requirements-build.txt \
+    --mount=type=bind,src=requirements/build.txt,target=requirements/build.txt \
     pip install --upgrade pip && \
-    pip install -r requirements-build.txt
+    pip install -r requirements/build.txt
 
 FROM cpu-test-arm AS build
 
 WORKDIR /workspace/vllm
 
 RUN --mount=type=cache,target=/root/.cache/pip \
-    --mount=type=bind,src=requirements-common.txt,target=requirements-common.txt \
+    --mount=type=bind,src=requirements/common.txt,target=requirements/common.txt \
-    --mount=type=bind,src=requirements-cpu.txt,target=requirements-cpu.txt \
+    --mount=type=bind,src=requirements/cpu.txt,target=requirements/cpu.txt \
-    pip install -v -r requirements-cpu.txt
+    pip install -v -r requirements/cpu.txt
 
 COPY . .
 ARG GIT_REPO_CHECK=0

Dockerfile.cpu

@@ -29,18 +29,18 @@ WORKDIR /workspace
 ARG PIP_EXTRA_INDEX_URL="https://download.pytorch.org/whl/cpu"
 ENV PIP_EXTRA_INDEX_URL=${PIP_EXTRA_INDEX_URL}
 RUN --mount=type=cache,target=/root/.cache/pip \
-    --mount=type=bind,src=requirements-build.txt,target=requirements-build.txt \
+    --mount=type=bind,src=requirements/build.txt,target=requirements/build.txt \
     pip install --upgrade pip && \
-    pip install -r requirements-build.txt
+    pip install -r requirements/build.txt
 
 FROM cpu-test-1 AS build
 
 WORKDIR /workspace/vllm
 
 RUN --mount=type=cache,target=/root/.cache/pip \
-    --mount=type=bind,src=requirements-common.txt,target=requirements-common.txt \
+    --mount=type=bind,src=requirements/common.txt,target=requirements/common.txt \
-    --mount=type=bind,src=requirements-cpu.txt,target=requirements-cpu.txt \
+    --mount=type=bind,src=requirements/cpu.txt,target=requirements/cpu.txt \
-    pip install -v -r requirements-cpu.txt
+    pip install -v -r requirements/cpu.txt
 
 COPY . .
 ARG GIT_REPO_CHECK=0

Dockerfile.hpu

@@ -4,7 +4,7 @@ COPY ./ /workspace/vllm
 
 WORKDIR /workspace/vllm
 
-RUN pip install -v -r requirements-hpu.txt
+RUN pip install -v -r requirements/hpu.txt
 
 ENV no_proxy=localhost,127.0.0.1
 ENV PT_HPU_ENABLE_LAZY_COLLECTIVES=true

Dockerfile.neuron

@@ -36,7 +36,7 @@ RUN --mount=type=bind,source=.git,target=.git \
 
 RUN python3 -m pip install -U \
         'cmake>=3.26' ninja packaging 'setuptools-scm>=8' wheel jinja2 \
-        -r requirements-neuron.txt
+        -r requirements/neuron.txt
 
 ENV VLLM_TARGET_DEVICE neuron
 RUN --mount=type=bind,source=.git,target=.git \

Dockerfile.openvino

@@ -16,7 +16,7 @@ RUN --mount=type=bind,source=.git,target=.git \
 
 RUN python3 -m pip install -U pip
 # install build requirements
-RUN PIP_EXTRA_INDEX_URL="https://download.pytorch.org/whl/cpu" python3 -m pip install -r /workspace/requirements-build.txt
+RUN PIP_EXTRA_INDEX_URL="https://download.pytorch.org/whl/cpu" python3 -m pip install -r /workspace/requirements/build.txt
 # build vLLM with OpenVINO backend
 RUN PIP_EXTRA_INDEX_URL="https://download.pytorch.org/whl/cpu" VLLM_TARGET_DEVICE="openvino" python3 -m pip install /workspace
 

Dockerfile.ppc64le

@@ -6,7 +6,7 @@ ENV PATH="/usr/local/cargo/bin:$PATH:/opt/conda/bin/"
 
 RUN apt-get update -y && apt-get install -y git wget kmod curl vim libnuma-dev libsndfile-dev libprotobuf-dev build-essential ffmpeg libsm6 libxext6 libgl1 libssl-dev 
 
-# Some packages in requirements-cpu are installed here
+# Some packages in requirements/cpu are installed here
 # IBM provides optimized packages for ppc64le processors in the open-ce project for mamba
 # Currently these may not be available for venv or pip directly
 RUN micromamba install -y -n base -c https://ftp.osuosl.org/pub/open-ce/1.11.0-p10/ -c defaults python=3.10 rust && micromamba clean --all --yes
@@ -21,7 +21,7 @@ RUN --mount=type=bind,source=.git,target=.git \
 RUN --mount=type=cache,target=/root/.cache/pip  \
     RUSTFLAGS='-L /opt/conda/lib' pip install -v --prefer-binary --extra-index-url https://repo.fury.io/mgiessing \
         'cmake>=3.26' ninja packaging 'setuptools-scm>=8' wheel jinja2 \
-        -r requirements-cpu.txt \
+        -r requirements/cpu.txt \
         xformers uvloop==0.20.0
 
 RUN --mount=type=bind,source=.git,target=.git \

Dockerfile.rocm

@@ -38,7 +38,7 @@ FROM fetch_vllm AS build_vllm
 ARG USE_CYTHON
 # Build vLLM
 RUN cd vllm \
-    && python3 -m pip install -r requirements-rocm.txt \
+    && python3 -m pip install -r requirements/rocm.txt \
     && python3 setup.py clean --all  \
     && if [ ${USE_CYTHON} -eq "1" ]; then python3 setup_cython.py build_ext --inplace; fi \
     && python3 setup.py bdist_wheel --dist-dir=dist
@@ -60,7 +60,7 @@ RUN python3 -m pip install --upgrade pip && rm -rf /var/lib/apt/lists/*
 # Install vLLM
 RUN --mount=type=bind,from=export_vllm,src=/,target=/install \
     cd /install \
-    && pip install -U -r requirements-rocm.txt \
+    && pip install -U -r requirements/rocm.txt \
     && pip uninstall -y vllm \
     && pip install *.whl
 
@@ -99,7 +99,7 @@ RUN if [ ${BUILD_RPD} -eq "1" ]; then \
 # Install vLLM
 RUN --mount=type=bind,from=export_vllm,src=/,target=/install \
     cd /install \
-    && pip install -U -r requirements-rocm.txt \
+    && pip install -U -r requirements/rocm.txt \
     && pip uninstall -y vllm \
     && pip install *.whl
 

Dockerfile.s390x

@@ -58,7 +58,7 @@ RUN --mount=type=cache,target=/root/.cache/uv \
     cd ../../python && \
     export PYARROW_PARALLEL=4 && \
     export ARROW_BUILD_TYPE=release && \
-    uv pip install -r requirements-build.txt && \
+    uv pip install -r requirements/build.txt && \
     python setup.py build_ext --build-type=$ARROW_BUILD_TYPE --bundle-arrow-cpp bdist_wheel
 
 FROM python-install AS numa-build
@@ -120,7 +120,7 @@ RUN --mount=type=cache,target=/root/.cache/uv \
     --mount=type=bind,from=rust,source=/root/.rustup,target=/root/.rustup,rw \
     --mount=type=bind,from=pyarrow,source=/tmp/arrow/python/dist,target=/tmp/arrow-wheels \
     --mount=type=bind,from=torch-vision,source=/tmp/vision/dist,target=/tmp/vision-wheels/ \
-     sed -i '/^torch/d' requirements-build.txt && \
+     sed -i '/^torch/d' requirements/build.txt && \
      ARROW_WHL_FILE=$(ls /tmp/arrow-wheels/pyarrow-*.whl | head -n 1) && \
      VISION_WHL_FILE=$(ls /tmp/vision-wheels/*.whl | head -n 1) && \
     uv pip install -v \    
@@ -128,8 +128,8 @@ RUN --mount=type=cache,target=/root/.cache/uv \
         $VISION_WHL_FILE \
         --extra-index-url https://download.pytorch.org/whl/nightly/cpu \
         --index-strategy unsafe-best-match \
-        -r requirements-build.txt \
+        -r requirements/build.txt \
-        -r requirements-cpu.txt 
+        -r requirements/cpu.txt 
 
 # Build and install vllm
 RUN --mount=type=cache,target=/root/.cache/uv \

Dockerfile.tpu

@@ -19,7 +19,7 @@ ENV VLLM_TARGET_DEVICE="tpu"
 RUN --mount=type=cache,target=/root/.cache/pip \
     --mount=type=bind,source=.git,target=.git \
     python3 -m pip install \
-        -r requirements-tpu.txt
+        -r requirements/tpu.txt
 RUN python3 setup.py develop
 
 # install development dependencies (for testing)

Dockerfile.xpu

@@ -25,12 +25,12 @@ RUN apt-get update -y && \
     wget
 
 WORKDIR /workspace/vllm
-COPY requirements-xpu.txt /workspace/vllm/requirements-xpu.txt
+COPY requirements/xpu.txt /workspace/vllm/requirements/xpu.txt
-COPY requirements-common.txt /workspace/vllm/requirements-common.txt
+COPY requirements/common.txt /workspace/vllm/requirements/common.txt
 
 RUN --mount=type=cache,target=/root/.cache/pip \
     pip install --no-cache-dir \
-    -r requirements-xpu.txt
+    -r requirements/xpu.txt
 
 RUN git clone https://github.com/intel/pti-gpu && \
     cd pti-gpu/sdk && \

MANIFEST.in

@@ -1,9 +1,9 @@
 include LICENSE
-include requirements-common.txt
+include requirements/common.txt
-include requirements-cuda.txt
+include requirements/cuda.txt
-include requirements-rocm.txt
+include requirements/rocm.txt
-include requirements-neuron.txt
+include requirements/neuron.txt
-include requirements-cpu.txt
+include requirements/cpu.txt
 include CMakeLists.txt
 
 recursive-include cmake *

docs/README.md

@@ -4,7 +4,7 @@
 
 ```bash
 # Install dependencies.
-pip install -r requirements-docs.txt
+pip install -r ../requirements/docs.txt
 
 # Build the docs.
 make clean

docs/source/contributing/overview.md

@@ -23,7 +23,7 @@ Check out the [building from source](#build-from-source) documentation for detai
 ## Testing
 
 ```bash
-pip install -r requirements-dev.txt
+pip install -r requirements/dev.txt
 
 # Linting, formatting and static type checking
 pre-commit install --hook-type pre-commit --hook-type commit-msg

docs/source/getting_started/installation/ai_accelerator/hpu-gaudi.inc.md

@@ -63,7 +63,7 @@ To build and install vLLM from source, run:
 ```console
 git clone https://github.com/vllm-project/vllm.git
 cd vllm
-pip install -r requirements-hpu.txt
+pip install -r requirements/hpu.txt
 python setup.py develop
 ```
 
@@ -73,7 +73,7 @@ Currently, the latest features and performance optimizations are developed in Ga
 git clone https://github.com/HabanaAI/vllm-fork.git
 cd vllm-fork
 git checkout habana_main
-pip install -r requirements-hpu.txt
+pip install -r requirements/hpu.txt
 python setup.py develop
 ```
 

docs/source/getting_started/installation/ai_accelerator/neuron.inc.md

@@ -116,7 +116,7 @@ Once neuronx-cc and transformers-neuronx packages are installed, we will be able
 ```console
 git clone https://github.com/vllm-project/vllm.git
 cd vllm
-pip install -U -r requirements-neuron.txt
+pip install -U -r requirements/neuron.txt
 VLLM_TARGET_DEVICE="neuron" pip install .
 ```
 

docs/source/getting_started/installation/ai_accelerator/openvino.inc.md

@@ -32,7 +32,7 @@ Second, clone vLLM and install prerequisites for the vLLM OpenVINO backend insta
 ```console
 git clone https://github.com/vllm-project/vllm.git
 cd vllm
-pip install -r requirements-build.txt --extra-index-url https://download.pytorch.org/whl/cpu
+pip install -r requirements/build.txt --extra-index-url https://download.pytorch.org/whl/cpu
 ```
 
 Finally, install vLLM with OpenVINO backend:

docs/source/getting_started/installation/ai_accelerator/tpu.inc.md

@@ -151,7 +151,7 @@ pip uninstall torch torch-xla -y
 Install build dependencies:
 
 ```bash
-pip install -r requirements-tpu.txt
+pip install -r requirements/tpu.txt
 sudo apt-get install libopenblas-base libopenmpi-dev libomp-dev
 ```
 

docs/source/getting_started/installation/cpu/apple.inc.md

@@ -25,7 +25,7 @@ After installation of XCode and the Command Line Tools, which include Apple Clan
 ```console
 git clone https://github.com/vllm-project/vllm.git
 cd vllm
-pip install -r requirements-cpu.txt
+pip install -r requirements/cpu.txt
 pip install -e . 
 ```
 

docs/source/getting_started/installation/cpu/build.inc.md

@@ -18,7 +18,7 @@ Third, install Python packages for vLLM CPU backend building:
 ```console
 pip install --upgrade pip
 pip install "cmake>=3.26" wheel packaging ninja "setuptools-scm>=8" numpy
-pip install -v -r requirements-cpu.txt --extra-index-url https://download.pytorch.org/whl/cpu
+pip install -v -r requirements/cpu.txt --extra-index-url https://download.pytorch.org/whl/cpu
 ```
 
 Finally, build and install vLLM CPU backend:

docs/source/getting_started/installation/gpu/cuda.inc.md

@@ -148,7 +148,7 @@ To build vLLM using an existing PyTorch installation:
 git clone https://github.com/vllm-project/vllm.git
 cd vllm
 python use_existing_torch.py
-pip install -r requirements-build.txt
+pip install -r requirements/build.txt
 pip install -e . --no-build-isolation
 ```
 

docs/source/getting_started/installation/gpu/rocm.inc.md

@@ -84,7 +84,7 @@ Currently, there are no pre-built ROCm wheels.
     # Install dependencies
     $ pip install --upgrade numba scipy huggingface-hub[cli,hf_transfer] setuptools_scm
     $ pip install "numpy<2"
-    $ pip install -r requirements-rocm.txt
+    $ pip install -r requirements/rocm.txt
 
     # Build vLLM for MI210/MI250/MI300.
     $ export PYTORCH_ROCM_ARCH="gfx90a;gfx942"

docs/source/getting_started/installation/gpu/xpu.inc.md

@@ -25,7 +25,7 @@ Currently, there are no pre-built XPU wheels.
 ```console
 source /opt/intel/oneapi/setvars.sh
 pip install --upgrade pip
-pip install -v -r requirements-xpu.txt
+pip install -v -r requirements/xpu.txt
 ```
 
 - Finally, build and install vLLM XPU backend:

format.sh

@@ -1,6 +1,6 @@
 #!/bin/bash
 
 echo "vLLM linting system has been moved from format.sh to pre-commit hook."
-echo "Please run 'pip install -r requirements-lint.txt', followed by"
+echo "Please run 'pip install -r requirements/lint.txt', followed by"
 echo "'pre-commit install --hook-type pre-commit --hook-type commit-msg' to install the pre-commit hook."
 echo "Then linters will run automatically before each commit."

pyproject.toml

@@ -1,5 +1,5 @@
 [build-system]
-# Should be mirrored in requirements-build.txt
+# Should be mirrored in requirements/build.txt
 requires = [
     "cmake>=3.26",
     "ninja",

requirements/cpu.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 # Dependencies for CPUs
 torch==2.5.1+cpu; platform_machine != "ppc64le" and platform_machine != "aarch64" and platform_system != "Darwin" and platform_machine != "s390x"

requirements/cuda.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 numba == 0.60.0 # v0.61 doesn't support Python 3.9. Required for N-gram speculative decoding
 

requirements/dev.txt

@@ -1,5 +1,5 @@
--r requirements-lint.txt
+-r lint.txt
--r requirements-test.txt
+-r test.txt
 
 # Avoid adding requirements directly to this file.
 # Instead, modify the two files referenced above.

requirements/hpu.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 # Dependencies for HPU code
 ray

requirements/neuron.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 # Dependencies for Neuron devices
 torch-neuronx >= 2.5.0

requirements/openvino.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 torch == 2.5.1 #  should be aligned with "common" vLLM torch version
 openvino >= 2024.4.0 # since 2024.4.0 both CPU and GPU support Paged Attention

requirements/rocm-build.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 --extra-index-url https://download.pytorch.org/whl/rocm6.2
 torch==2.5.1

requirements/rocm.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 numba == 0.60.0 # v0.61 doesn't support Python 3.9. Required for N-gram speculative decoding
 

requirements/test.txt

@@ -1,5 +1,5 @@
 # This file was autogenerated by uv via the following command:
-#    uv pip compile requirements-test.in -o requirements-test.txt
+#    uv pip compile requirements/test.in -o requirements/test.txt
 absl-py==2.1.0
     # via rouge-score
 accelerate==1.0.1
@@ -32,9 +32,9 @@ attrs==24.2.0
 audioread==3.0.1
     # via librosa
 awscli==1.35.23
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 bitsandbytes==0.45.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 black==24.10.0
     # via datamodel-code-generator
 boto3==1.35.57
@@ -47,7 +47,7 @@ botocore==1.35.57
 bounded-pool-executor==0.0.3
     # via pqdm
 buildkite-test-collector==0.1.9
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 certifi==2024.8.30
     # via
     #   httpcore
@@ -79,7 +79,7 @@ cupy-cuda12x==13.3.0
 cycler==0.12.1
     # via matplotlib
 datamodel-code-generator==0.26.3
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 dataproperty==1.0.1
     # via
     #   pytablewriter
@@ -91,7 +91,7 @@ datasets==3.0.2
 decorator==5.1.1
     # via librosa
 decord==0.6.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 dill==0.3.8
     # via
     #   datasets
@@ -104,7 +104,7 @@ docutils==0.16
     # via awscli
 einops==0.8.0
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   encodec
     #   vector-quantize-pytorch
     #   vocos
@@ -145,7 +145,7 @@ fsspec==2024.9.0
     #   huggingface-hub
     #   torch
 genai-perf==0.0.8
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 genson==1.3.0
     # via datamodel-code-generator
 h11==0.14.0
@@ -155,7 +155,7 @@ hiredis==3.0.0
 httpcore==1.0.6
     # via httpx
 httpx==0.27.2
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 huggingface-hub==0.26.2
     # via
     #   accelerate
@@ -187,7 +187,7 @@ jinja2==3.1.4
     #   datamodel-code-generator
     #   torch
 jiwer==3.0.5
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 jmespath==1.0.1
     # via
     #   boto3
@@ -214,11 +214,11 @@ lazy-loader==0.4
 libnacl==2.1.0
     # via tensorizer
 librosa==0.10.2.post1
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 llvmlite==0.43.0
     # via numba
 lm-eval==0.4.4
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 lxml==5.3.0
     # via sacrebleu
 markdown-it-py==3.0.0
@@ -226,7 +226,7 @@ markdown-it-py==3.0.0
 markupsafe==3.0.2
     # via jinja2
 matplotlib==3.9.2
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 mbstrdecoder==1.1.3
     # via
     #   dataproperty
@@ -235,7 +235,7 @@ mbstrdecoder==1.1.3
 mdurl==0.1.2
     # via markdown-it-py
 mistral-common==1.5.1
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 more-itertools==10.5.0
     # via lm-eval
 mpmath==1.3.0
@@ -264,7 +264,7 @@ numexpr==2.10.1
     # via lm-eval
 numpy==1.26.4
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   accelerate
     #   bitsandbytes
     #   contourpy
@@ -366,7 +366,7 @@ patsy==1.0.1
     # via statsmodels
 peft==0.13.2
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   lm-eval
 pillow==10.4.0
     # via
@@ -388,7 +388,7 @@ pooch==1.8.2
 portalocker==2.10.1
     # via sacrebleu
 pqdm==0.2.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 propcache==0.2.0
     # via yarl
 protobuf==5.28.3
@@ -426,7 +426,7 @@ pytablewriter==1.2.0
     # via lm-eval
 pytest==8.3.3
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   buildkite-test-collector
     #   genai-perf
     #   pytest-asyncio
@@ -435,15 +435,15 @@ pytest==8.3.3
     #   pytest-rerunfailures
     #   pytest-shard
 pytest-asyncio==0.24.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 pytest-forked==1.6.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 pytest-mock==3.14.0
     # via genai-perf
 pytest-rerunfailures==14.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 pytest-shard==0.1.2
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 python-dateutil==2.9.0.post0
     # via
     #   botocore
@@ -473,7 +473,7 @@ pyyaml==6.0.2
 rapidfuzz==3.12.1
     # via jiwer
 ray==2.43.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 redis==5.2.0
     # via tensorizer
 referencing==0.35.1
@@ -512,9 +512,9 @@ rpds-py==0.20.1
 rsa==4.7.2
     # via awscli
 runai-model-streamer==0.11.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 runai-model-streamer-s3==0.11.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 s3transfer==0.10.3
     # via
     #   awscli
@@ -540,7 +540,7 @@ scipy==1.13.1
     #   statsmodels
     #   vocos
 sentence-transformers==3.2.1
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 sentencepiece==0.2.0
     # via mistral-common
 setuptools==75.8.0
@@ -557,7 +557,7 @@ sniffio==1.3.1
     #   httpx
 soundfile==0.12.1
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   librosa
 soxr==0.5.0.post1
     # via librosa
@@ -580,7 +580,7 @@ tenacity==9.0.0
     #   lm-eval
     #   plotly
 tensorizer==2.9.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 threadpoolctl==3.5.0
     # via scikit-learn
 tiktoken==0.7.0
@@ -588,12 +588,12 @@ tiktoken==0.7.0
     #   lm-eval
     #   mistral-common
 timm==1.0.11
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 tokenizers==0.21.0
     # via transformers
 torch==2.5.1
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   accelerate
     #   bitsandbytes
     #   encodec
@@ -609,7 +609,7 @@ torch==2.5.1
     #   vocos
 torchaudio==2.5.1
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   encodec
     #   vocos
 torchvision==0.20.1
@@ -630,19 +630,19 @@ tqdm-multiprocess==0.0.11
     # via lm-eval
 transformers==4.48.2
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   genai-perf
     #   lm-eval
     #   peft
     #   sentence-transformers
     #   transformers-stream-generator
 transformers-stream-generator==0.0.5
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 triton==3.1.0
     # via torch
 tritonclient==2.51.0
     # via
-    #   -r requirements-test.in
+    #   -r requirements/test.in
     #   genai-perf
 typepy==1.3.2
     # via
@@ -668,9 +668,9 @@ urllib3==2.2.3
     #   responses
     #   tritonclient
 vector-quantize-pytorch==1.21.2
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 vocos==0.1.0
-    # via -r requirements-test.in
+    # via -r requirements/test.in
 word2number==1.1
     # via lm-eval
 xxhash==3.5.0

requirements/tpu.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 # Dependencies for TPU
 cmake>=3.26

requirements/xpu.txt

@@ -1,5 +1,5 @@
 # Common dependencies
--r requirements-common.txt
+-r common.txt
 
 ray >= 2.9
 cmake>=3.26

setup.py

@@ -27,7 +27,7 @@ def load_module_from_path(module_name, path):
     return module
 
 
-ROOT_DIR = os.path.dirname(__file__)
+ROOT_DIR = Path(__file__).parent
 logger = logging.getLogger(__name__)
 
 # cannot import envs directly because it depends on vllm,
@@ -520,10 +520,6 @@ def get_nvcc_cuda_version() -> Version:
     return nvcc_cuda_version
 
 
-def get_path(*filepath) -> str:
-    return os.path.join(ROOT_DIR, *filepath)
-
-
 def get_gaudi_sw_version():
     """
     Returns the driver version.
@@ -591,9 +587,10 @@ def get_vllm_version() -> str:
 
 def get_requirements() -> list[str]:
     """Get Python package dependencies from requirements.txt."""
+    requirements_dir = ROOT_DIR / "requirements"
 
     def _read_requirements(filename: str) -> list[str]:
-        with open(get_path(filename)) as f:
+        with open(requirements_dir / filename) as f:
             requirements = f.read().strip().split("\n")
         resolved_requirements = []
         for line in requirements:
@@ -606,9 +603,9 @@ def get_requirements() -> list[str]:
         return resolved_requirements
 
     if _no_device():
-        requirements = _read_requirements("requirements-common.txt")
+        requirements = _read_requirements("common.txt")
     elif _is_cuda():
-        requirements = _read_requirements("requirements-cuda.txt")
+        requirements = _read_requirements("cuda.txt")
         cuda_major, cuda_minor = torch.version.cuda.split(".")
         modified_requirements = []
         for req in requirements:
@@ -619,19 +616,19 @@ def get_requirements() -> list[str]:
             modified_requirements.append(req)
         requirements = modified_requirements
     elif _is_hip():
-        requirements = _read_requirements("requirements-rocm.txt")
+        requirements = _read_requirements("rocm.txt")
     elif _is_neuron():
-        requirements = _read_requirements("requirements-neuron.txt")
+        requirements = _read_requirements("neuron.txt")
     elif _is_hpu():
-        requirements = _read_requirements("requirements-hpu.txt")
+        requirements = _read_requirements("hpu.txt")
     elif _is_openvino():
-        requirements = _read_requirements("requirements-openvino.txt")
+        requirements = _read_requirements("openvino.txt")
     elif _is_tpu():
-        requirements = _read_requirements("requirements-tpu.txt")
+        requirements = _read_requirements("tpu.txt")
     elif _is_cpu():
-        requirements = _read_requirements("requirements-cpu.txt")
+        requirements = _read_requirements("cpu.txt")
     elif _is_xpu():
-        requirements = _read_requirements("requirements-xpu.txt")
+        requirements = _read_requirements("xpu.txt")
     else:
         raise ValueError(
             "Unsupported platform, please use CUDA, ROCm, Neuron, HPU, "

use_existing_torch.py

@@ -2,7 +2,7 @@
 
 import glob
 
-requires_files = glob.glob('requirements*.txt')
+requires_files = glob.glob('requirements/*.txt')
 requires_files += ["pyproject.toml"]
 for file in requires_files:
     print(f">>> cleaning {file}")