Merge pull request #651 from allenai/shanea/lumi-torch2.3-2

Update LUMI Dockerfile
allenai · Jul 11, 2024 · b10ab4b · b10ab4b
2 parents a101b31 + e0efc20
commit b10ab4b
Showing 1 changed file with 22 additions and 15 deletions.
diff --git a/docker/Dockerfile.lumi b/docker/Dockerfile.lumi
@@ -1,4 +1,4 @@
-FROM ubuntu:latest
+FROM ubuntu:jammy
 
 ENV DEBIAN_FRONTEND=noninteractive
 ENV LC_ALL=C.UTF-8
@@ -21,9 +21,9 @@ RUN apt-get install -y \
   vim \
   fish \
   wget \
+  unzip \
   parallel \
   s3cmd \
-  awscli \
   htop \
   wget \
   fish \
@@ -38,17 +38,22 @@ RUN apt-get install -y \
   gdb \
   apt-utils
 
+RUN curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
+RUN unzip awscliv2.zip
+RUN ./aws/install
+RUN rm -r ./aws/
+
 # Install Google tools
 RUN echo "deb [signed-by=/usr/share/keyrings/cloud.google.gpg] https://packages.cloud.google.com/apt cloud-sdk main" | tee -a /etc/apt/sources.list.d/google-cloud-sdk.list
 RUN curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key --keyring /usr/share/keyrings/cloud.google.gpg add -
 RUN apt-get update
-RUN apt-get install google-cloud-cli
+RUN apt-get install -y google-cloud-cli
 
 # Install ROCm
 RUN mkdir --parents --mode=0755 /etc/apt/keyrings && \
     wget https://repo.radeon.com/rocm/rocm.gpg.key -O - | gpg --dearmor | tee /etc/apt/keyrings/rocm.gpg > /dev/null && \
-    echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/amdgpu/5.6/ubuntu jammy main" >> /etc/apt/sources.list.d/amdgpu.list && \
-    echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/rocm/apt/5.6 jammy main" >> /etc/apt/sources.list.d/rocm.list && \
+    echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/amdgpu/6.0.2/ubuntu jammy main" >> /etc/apt/sources.list.d/amdgpu.list && \
+    echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/rocm/apt/6.0.2 jammy main" >> /etc/apt/sources.list.d/rocm.list && \
     echo 'Package: *\nPin: release o=repo.radeon.com\nPin-Priority: 600' > /etc/apt/preferences.d/rocm-pin-600 && \
     apt-get update && \
     apt-get install -y rccl rccl-dev rocm-libs rocm-gdb rocm-dev rocm-developer-tools rocm-hip-runtime-dev rocm-utils rocm-hip-sdk && \
@@ -83,7 +88,7 @@ ENV LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH
 
 # Install torch
 RUN pip install --upgrade pip
-RUN pip install torch==2.1.1 torchvision==0.16.1 torchaudio==2.1.1 --index-url https://download.pytorch.org/whl/rocm5.6
+RUN pip install --no-cache-dir torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/rocm6.0
 
 # Install DeepSpeed
 RUN pip install --no-cache-dir mpi4py
@@ -95,22 +100,24 @@ RUN cd /opt && \
 
 # Install more dependencies
 COPY pyproject.toml .
-RUN mkdir olmo && touch olmo/__init__.py && \
-    pip install --no-cache-dir .[all] && \
+RUN mkdir olmo && \
+    touch olmo/__init__.py && \
+    echo 'VERSION = "0.1.0"' > olmo/version.py && \
+    pip install --no-cache-dir .[train] && \
     pip uninstall -y ai2-olmo && \
     rm -rf olmo/
 
 RUN pip install --no-cache-dir py-spy
 RUN pip install --no-cache-dir wandb --upgrade
 
+# # Install flash attention (for MI200 series!)
+# RUN cd /opt && \
+#     git clone --recursive https://github.com/ROCm/flash-attention.git && \
+#     cd flash-attention && \
+#     GPU_ARCHS="gfx90a" pip install .
+
 # Cleanup
 RUN apt-get autoremove
 RUN rm -rf /opt/mpich-3.1.4 /opt/aws-ofi-rccl /opt/DeepSpeed
 RUN apt-get clean
-RUN pip cache purge
-
-# Install flash attention (for MI200 series!)
-RUN cd /opt && \
-    git clone --recursive https://github.com/ROCm/flash-attention.git && \
-    cd flash-attention && \
-    GPU_ARCHS="gfx90a" pip install .
+RUN pip cache purge