MNT Remove Python 3.8 since it's end of life (#2135)

The end of life of Python 3.8 has arrived:

https://devguide.python.org/versions/

Therefore, Python 3.8 is removed from CI.

By default, Python 3.11 is now used.

Python 3.12 should be added to the CI matrix now, but that's for a
separate PR.

Also fixed:

The workflow tried to build on top of docker/README.md because globbing
was too broadly defined.

Reduce unnecessary steps to hopefully get disk space usage down, as
GitHub action currently fails with not enough disk space.

.github/workflows/stale.yml

@@ -20,7 +20,7 @@ jobs:
     - name: Setup Python
       uses: actions/setup-python@v4
       with:
-        python-version: 3.8
+        python-version: 3.11
 
     - name: Install requirements
       run: |

.github/workflows/test-docker-build.yml

@@ -4,7 +4,7 @@ on:
   pull_request:
     paths:
       # Run only when DockerFile files are modified
-      - "docker/**"
+      - "docker/*/Dockerfile"
 jobs:
   get_changed_files:
     name: "Build all modified docker images"
@@ -18,7 +18,7 @@ jobs:
         id: changed-files
         uses: tj-actions/changed-files@1c8e6069583811afb28f97afeaf8e7da80c6be5c #v42
         with:
-          files: docker/**
+          files: docker/*/Dockerfile
           json: "true"
       - name: Run step if only the files listed above change
         if: steps.changed-files.outputs.any_changed == 'true'

.github/workflows/tests.yml

@@ -17,7 +17,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v4
         with:
-          python-version: "3.8"
+          python-version: "3.11"
           cache: "pip"
           cache-dependency-path: "setup.py"
       - name: Install dependencies
@@ -34,7 +34,7 @@ jobs:
       # TODO: remove 'fail-fast' line once timeout issue from the Hub is solved
       fail-fast: false
       matrix:
-        python-version: ["3.8", "3.9", "3.10", "3.11"]
+        python-version: ["3.9", "3.10", "3.11"]
         os: ["ubuntu-latest", "macos-12", "windows-latest"]
     runs-on: ${{ matrix.os }}
     steps:

docker/README.md

@@ -1,6 +1,6 @@
 # PEFT Docker images
 
-Here we store all PEFT Docker images used in our testing infrastructure. We use python 3.8 for now on all our images.
+Here we store all PEFT Docker images used in our testing infrastructure. We use python 3.11 for now on all our images.
 
 - `peft-cpu`: PEFT compiled on CPU with all other HF libraries installed on main branch
 - `peft-gpu`: PEFT complied for NVIDIA GPUs wih all other HF libraries installed on main branch

docker/peft-cpu/Dockerfile

@@ -4,7 +4,7 @@
 # Use base conda image to reduce time
 FROM continuumio/miniconda3:latest AS compile-image
 # Specify py version
-ENV PYTHON_VERSION=3.8
+ENV PYTHON_VERSION=3.11
 # Install apt libs - copied from https://github.com/huggingface/accelerate/blob/main/docker/accelerate-gpu/Dockerfile
 RUN apt-get update && \
     apt-get install -y curl git wget software-properties-common git-lfs && \

docker/peft-gpu-bnb-latest/Dockerfile

@@ -4,7 +4,7 @@
 # Use base conda image to reduce time
 FROM continuumio/miniconda3:latest AS compile-image
 # Specify py version
-ENV PYTHON_VERSION=3.8
+ENV PYTHON_VERSION=3.11
 # Install apt libs - copied from https://github.com/huggingface/accelerate/blob/main/docker/accelerate-gpu/Dockerfile
 RUN apt-get update && \
     apt-get install -y curl git wget software-properties-common git-lfs && \

docker/peft-gpu-bnb-multi-source/Dockerfile

@@ -4,7 +4,7 @@
 # Use base conda image to reduce time
 FROM continuumio/miniconda3:latest AS compile-image
 # Specify py version
-ENV PYTHON_VERSION=3.8
+ENV PYTHON_VERSION=3.11
 # Install apt libs - copied from https://github.com/huggingface/accelerate/blob/main/docker/accelerate-gpu/Dockerfile
 RUN apt-get update && \
     apt-get install -y curl git wget software-properties-common git-lfs && \

docker/peft-gpu-bnb-source/Dockerfile

@@ -4,7 +4,7 @@
 # Use base conda image to reduce time
 FROM continuumio/miniconda3:latest AS compile-image
 # Specify py version
-ENV PYTHON_VERSION=3.8
+ENV PYTHON_VERSION=3.11
 # Install apt libs - copied from https://github.com/huggingface/accelerate/blob/main/docker/accelerate-gpu/Dockerfile
 RUN apt-get update && \
     apt-get install -y curl git wget software-properties-common git-lfs && \

docker/peft-gpu/Dockerfile

@@ -4,23 +4,18 @@
 # Use base conda image to reduce time
 FROM continuumio/miniconda3:latest AS compile-image
 # Specify py version
-ENV PYTHON_VERSION=3.8
+ENV PYTHON_VERSION=3.11
 # Install apt libs - copied from https://github.com/huggingface/accelerate/blob/main/docker/accelerate-gpu/Dockerfile
+# Install audio-related libraries
 RUN apt-get update && \
-    apt-get install -y curl git wget software-properties-common git-lfs && \
+    apt-get install -y curl git wget software-properties-common git-lfs ffmpeg libsndfile1-dev && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists*
 
-# Install audio-related libraries 
-RUN apt-get update && \
-    apt install -y ffmpeg
-
-RUN apt install -y libsndfile1-dev
 RUN git lfs install
 
 # Create our conda env - copied from https://github.com/huggingface/accelerate/blob/main/docker/accelerate-gpu/Dockerfile
 RUN conda create --name peft python=${PYTHON_VERSION} ipython jupyter pip
-RUN python3 -m pip install --no-cache-dir --upgrade pip
 
 # Below is copied from https://github.com/huggingface/accelerate/blob/main/docker/accelerate-gpu/Dockerfile
 # We don't install pytorch here yet since CUDA isn't available
@@ -35,25 +30,20 @@ FROM nvidia/cuda:12.2.2-devel-ubuntu22.04 AS build-image
 COPY --from=compile-image /opt/conda /opt/conda
 ENV PATH /opt/conda/bin:$PATH
 
-RUN chsh -s /bin/bash
-SHELL ["/bin/bash", "-c"]
-RUN source activate peft && \ 
-    python3 -m pip install --no-cache-dir bitsandbytes optimum auto-gptq
-
-# Add autoawq for quantization testing
-RUN source activate peft && \
-    python3 -m pip install --no-cache-dir https://github.com/casper-hansen/AutoAWQ/releases/download/v0.2.4/autoawq-0.2.4-cp38-cp38-linux_x86_64.whl
-RUN source activate peft && \
-    python3 -m pip install --no-cache-dir https://github.com/casper-hansen/AutoAWQ_kernels/releases/download/v0.0.6/autoawq_kernels-0.0.6-cp38-cp38-linux_x86_64.whl
-
 # Install apt libs
 RUN apt-get update && \
     apt-get install -y curl git wget && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists*
 
-# Add eetq for quantization testing
-RUN source activate peft && \
+RUN chsh -s /bin/bash
+SHELL ["/bin/bash", "-c"]
+RUN source activate peft && \ 
+    python3 -m pip install --no-cache-dir bitsandbytes optimum auto-gptq && \
+    # Add autoawq for quantization testing
+    python3 -m pip install --no-cache-dir https://github.com/casper-hansen/AutoAWQ/releases/download/v0.2.6/autoawq-0.2.6-cp311-cp311-linux_x86_64.whl && \
+    python3 -m pip install --no-cache-dir https://github.com/casper-hansen/AutoAWQ_kernels/releases/download/v0.0.8/autoawq_kernels-0.0.8-cp311-cp311-linux_x86_64.whl && \
+    # Add eetq for quantization testing
     python3 -m pip install git+https://github.com/NetEase-FuXi/EETQ.git
 
 # Activate the conda env and install transformers + accelerate from source
@@ -65,17 +55,13 @@ RUN source activate peft && \
     torchao \
     git+https://github.com/huggingface/transformers \
     git+https://github.com/huggingface/accelerate \
-    peft[test]@git+https://github.com/huggingface/peft
+    peft[test]@git+https://github.com/huggingface/peft \
+    # Add aqlm for quantization testing
+    aqlm[gpu]>=1.0.2 \
+    # Add HQQ for quantization testing
+    hqq
 
-# Add aqlm for quantization testing
 RUN source activate peft && \
-    pip install aqlm[gpu]>=1.0.2
-
-# Add HQQ for quantization testing
-RUN source activate peft && \
-pip install hqq
-
-RUN source activate peft && \ 
     pip freeze | grep transformers
 
 RUN echo "source activate peft" >> ~/.profile

docs/source/install.md

@@ -16,7 +16,7 @@ rendered properly in your Markdown viewer.
 
 # Installation
 
-Before you start, you will need to setup your environment, install the appropriate packages, and configure 🤗 PEFT. 🤗 PEFT is tested on **Python 3.8+**.
+Before you start, you will need to setup your environment, install the appropriate packages, and configure 🤗 PEFT. 🤗 PEFT is tested on **Python 3.9+**.
 
 🤗 PEFT is available on PyPI, as well as GitHub:
 

setup.py

@@ -54,7 +54,7 @@ setup(
     packages=find_packages("src"),
     package_data={"peft": ["py.typed", "tuners/boft/fbd/fbd_cuda.cpp", "tuners/boft/fbd/fbd_cuda_kernel.cu"]},
     entry_points={},
-    python_requires=">=3.8.0",
+    python_requires=">=3.9.0",
     install_requires=[
         "numpy>=1.17",
         "packaging>=20.0",
@@ -76,7 +76,7 @@ setup(
         "License :: OSI Approved :: Apache Software License",
         "Operating System :: OS Independent",
         "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.8",
+        "Programming Language :: Python :: 3.11",
         "Topic :: Scientific/Engineering :: Artificial Intelligence",
     ],
 )

tests/test_gpu_examples.py

@@ -2963,10 +2963,8 @@ class PeftHqqGPUTests(unittest.TestCase):
         assert cc_matrix.min() > 0.97
 
 
-# TODO: unskip the tests once https://github.com/casper-hansen/AutoAWQ/issues/466 is fixed
 @require_torch_gpu
 @require_auto_awq
-@pytest.mark.skip(reason="Needs https://github.com/casper-hansen/AutoAWQ/issues/466 to be fixed first")
 class PeftAwqGPUTests(unittest.TestCase):
     r"""
     Awq + peft tests