Update Docker image builds for torch 2.7+cu126 (#2514)

* Update Docker image builds for torch 2.7+cu126

* Remove bnb multi-source dockerfile

Author: matthewdouglas

.github/workflows/build_docker_images.yml

@@ -148,37 +148,3 @@ jobs:
           status: ${{ job.status }}
           slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
 
-  latest-cuda-bnb-source-multi:
-    name: "Latest Peft GPU + bnb (multi-backend) source [accelerate / peft / transformers source]"
-    runs-on:
-      group: aws-general-8-plus
-    steps:
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      - name: Check out code
-        uses: actions/checkout@v3
-        with:
-          persist-credentials: false
-      - name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-
-      - name: Build and Push GPU
-        uses: docker/build-push-action@v4
-        with:
-          context: ./docker/peft-gpu-bnb-multi-source
-          push: true
-          tags: huggingface/peft-gpu-bnb-multi-source
-
-      - name: Post to Slack
-        if: always()
-        uses: huggingface/hf-workflows/.github/actions/post-slack@main
-        with:
-          slack_channel: ${{ env.CI_SLACK_CHANNEL }}
-          title: 🤗 Results of the PEFT-GPU (bnb source multi-backend / HF latest) docker build
-          status: ${{ job.status }}
-          slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
-
-

.github/workflows/nightly-bnb.yml

@@ -20,7 +20,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-          docker-image-name: ["huggingface/peft-gpu-bnb-source:latest", "huggingface/peft-gpu-bnb-latest:latest", "huggingface/peft-gpu-bnb-multi-source:latest"]
+          docker-image-name: ["huggingface/peft-gpu-bnb-source:latest", "huggingface/peft-gpu-bnb-latest:latest"]
     runs-on:
       group: aws-g6-4xlarge-plus
     env:
@@ -145,7 +145,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        docker-image-name: ["huggingface/peft-gpu-bnb-source:latest", "huggingface/peft-gpu-bnb-latest:latest", "huggingface/peft-gpu-bnb-multi-source:latest"]
+        docker-image-name: ["huggingface/peft-gpu-bnb-source:latest", "huggingface/peft-gpu-bnb-latest:latest"]
     runs-on:
       group: aws-g6-12xlarge-plus
     env:

docker/README.md

@@ -6,6 +6,3 @@ Here we store all PEFT Docker images used in our testing infrastructure. We use
 - `peft-gpu`: PEFT complied for NVIDIA GPUs with all other HF libraries installed on main branch
 - `peft-gpu-bnb-source`: PEFT complied for NVIDIA GPUs with `bitsandbytes` and all other HF libraries installed from main branch
 - `peft-gpu-bnb-latest`: PEFT complied for NVIDIA GPUs with `bitsandbytes` complied from main and all other HF libraries installed from latest PyPi
-- `peft-gpu-bnb-multi-source`: PEFT complied for NVIDIA GPUs with `bitsandbytes` complied from `multi-backend` branch and all other HF libraries installed from main branch
-
-`peft-gpu-bnb-source` and `peft-gpu-bnb-multi-source` are essentially the same, with the only difference being `bitsandbytes` compiled on another branch. Make sure to propagate the changes you applied on one file to the other!

docker/peft-gpu-bnb-latest/Dockerfile

@@ -31,7 +31,7 @@ RUN chsh -s /bin/bash
 SHELL ["/bin/bash", "-c"]
 
 # Stage 2
-FROM nvidia/cuda:12.4.1-devel-ubuntu22.04 AS build-image
+FROM nvidia/cuda:12.6.3-devel-ubuntu22.04 AS build-image
 COPY --from=compile-image /opt/conda /opt/conda
 ENV PATH /opt/conda/bin:$PATH
 
@@ -56,7 +56,7 @@ RUN source activate peft && \
     peft \
     optimum \
     auto-gptq && \
-    git clone https://github.com/TimDettmers/bitsandbytes && cd bitsandbytes && \
+    git clone https://github.com/bitsandbytes-foundation/bitsandbytes && cd bitsandbytes && \
     cmake -B . -DCOMPUTE_BACKEND=cuda -S . && \
     cmake --build . && \
     pip install -e . && \ 

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

@@ -31,7 +31,7 @@ RUN chsh -s /bin/bash
 SHELL ["/bin/bash", "-c"]
 
 # Stage 2
-FROM nvidia/cuda:12.4.1-devel-ubuntu22.04 AS build-image
+FROM nvidia/cuda:12.6.3-devel-ubuntu22.04 AS build-image
 COPY --from=compile-image /opt/conda /opt/conda
 ENV PATH /opt/conda/bin:$PATH
 
@@ -56,7 +56,7 @@ RUN source activate peft && \
     peft[test]@git+https://github.com/huggingface/peft \
     optimum \
     auto-gptq && \
-    git clone https://github.com/TimDettmers/bitsandbytes && cd bitsandbytes && \
+    git clone https://github.com/bitsandbytes-foundation/bitsandbytes && cd bitsandbytes && \
     cmake -B . -DCOMPUTE_BACKEND=cuda -S . && \
     cmake --build . && \
     pip install -e . && \