[Workspace Tempaltes] Upgrade finetuning template (#41752)

* Update to torch 2.1.1 with cuda toolkit 11.8 Signed-off-by: Artur Niederfahrenhorst <[email protected]> * Fix dockerfile and requirements Signed-off-by: Artur Niederfahrenhorst <[email protected]> * pin peft Signed-off-by: Artur Niederfahrenhorst <[email protected]> * Change Dockerfile version number * change CUDA version to 121 --------- Signed-off-by: Artur Niederfahrenhorst <[email protected]>
ray-project · Dec 13, 2023 · 9562078 · 9562078
1 parent 32663c8
commit 9562078
Show file tree

Hide file tree

Showing 4 changed files with 9 additions and 8 deletions.
diff --git a/doc/source/templates/testing/cluster_envs/04_finetuning_llms_with_deepspeed.yaml b/doc/source/templates/testing/cluster_envs/04_finetuning_llms_with_deepspeed.yaml
@@ -7,8 +7,8 @@ debian_packages:
 
 python:
  pip_packages: [
- torch==2.0.0,
  torchvision==0.15.1,
+ peft==0.7.0,
  torchaudio==2.0.1,
  deepspeed,
  fairscale,
@@ -28,6 +28,8 @@ python:
  conda_packages: []
 
 post_build_cmds:
+ - pip uninstall bitsandbytes -y || true
+ - pip install torch==2.1.1 --index-url https://download.pytorch.org/whl/cu118
  # Install Ray
  - pip3 uninstall -y ray || true && pip3 install -U {{ env["RAY_WHEELS"] | default("ray") }}
  - {{ env["RAY_WHEELS_SANITY_CHECK"] | default("echo No Ray wheels sanity check") }}
diff --git a/doc/source/templates/testing/docker/04_finetuning_llms_with_deepspeed/Dockerfile b/doc/source/templates/testing/docker/04_finetuning_llms_with_deepspeed/Dockerfile
@@ -1,12 +1,15 @@
 # Dockerfile used to create the docker image for `04_finetuning_llms_with_deepspeed`.
-FROM anyscale/ray:2.7.1-py310-cu121
+FROM anyscale/ray:2.8.1-py310-cu121
 
 COPY requirements.txt ./
 
 RUN sudo apt-get update
 RUN sudo apt-get install -y libaio1
 
+RUN pip install --upgrade pip
 RUN pip install -U -r requirements.txt
+RUN pip install torch==2.1.1 --index-url https://download.pytorch.org/whl/cu121
+RUN pip uninstall bitsandbytes -y
 
 RUN echo "Testing Ray Import..." && python -c "import ray"
 RUN ray --version
diff --git a/doc/source/templates/testing/docker/04_finetuning_llms_with_deepspeed/requirements.txt b/doc/source/templates/testing/docker/04_finetuning_llms_with_deepspeed/requirements.txt
@@ -1,18 +1,14 @@
-torch==2.0.0
-torchvision==0.15.1
-torchaudio==2.0.1
 deepspeed
 fairscale
 transformers>=4.31.0
 dataset
 accelerate
 evaluate
-bitsandbytes
 wandb
 pytorch-lightning
 protobuf<3.21.0
 torchmetrics
 lm_eval==0.3.0
 tiktoken==0.1.2
 sentencepiece
-peft @ git+https://github.com/huggingface/peft.git@08368a1fba16de09756f067637ff326c71598fb3
+peft==0.7.0
diff --git a/release/ray_release/byod/byod_finetune_llvms.sh b/release/ray_release/byod/byod_finetune_llvms.sh
@@ -5,7 +5,7 @@
 set -exo pipefail
 
 pip3 install -U \
- torch==2.0.0 \
+ torch==2.1.1 \
  torchvision==0.15.1 \
  torchaudio==2.0.1 \
  deepspeed==0.12.3 \