chore: Release (#65)

Signed-off-by: Ce Gao <cegao@tensorchord.ai>
tensorchord · Jun 6, 2023 · 2a88a44 · 2a88a44
1 parent 26820c8
commit 2a88a44
Show file tree

Hide file tree

Showing 5 changed files with 82 additions and 6 deletions.
diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
@@ -24,7 +24,9 @@ jobs:
       matrix:
         image:
           - name: modelzai/llm-chatglm-6b-int4
-            dockerfile: ./images/base/Dockerfile
+            dockerfile: ./images/chatglm-6b-int4/Dockerfile
+          - name: modelzai/llm-bloomz-560m
+            dockerfile: ./images/bloomz-560m/Dockerfile
           - name: modelzai/llm-chatglm-6b
             dockerfile: ./images/chatglm-6b/Dockerfile
           - name: modelzai/llm-llama-7b

diff --git a/.github/workflows/gcr.yml b/.github/workflows/gcr.yml
@@ -24,7 +24,7 @@ jobs:
       matrix:
         image:
           - name: modelzai/llm-chatglm-6b-int4
-            dockerfile: ./images/base/Dockerfile
+            dockerfile: ./images/chatglm-6b-int4/Dockerfile
           - name: modelzai/llm-chatglm-6b
             dockerfile: ./images/chatglm-6b/Dockerfile
           - name: modelzai/llm-llama-7b

diff --git a/images/bloomz-560m/Dockerfile b/images/bloomz-560m/Dockerfile
@@ -0,0 +1,74 @@
+ARG base=nvidia/cuda:11.6.2-cudnn8-runtime-ubuntu20.04
+
+FROM ${base}
+
+ENV DEBIAN_FRONTEND=noninteractive LANG=en_US.UTF-8 LC_ALL=en_US.UTF-8
+ENV PATH /opt/conda/bin:$PATH
+
+ARG MOSEC_PORT=8080
+ENV MOSEC_PORT=${MOSEC_PORT}
+
+ARG CONDA_VERSION=py310_22.11.1-1
+
+RUN apt update && \
+    apt install -y --no-install-recommends \
+        wget \
+        git \
+        build-essential \
+        ca-certificates && \
+    rm -rf /var/lib/apt/lists/*
+
+RUN set -x && \
+    UNAME_M="$(uname -m)" && \
+    if [ "${UNAME_M}" = "x86_64" ]; then \
+        MINICONDA_URL="https://repo.anaconda.com/miniconda/Miniconda3-${CONDA_VERSION}-Linux-x86_64.sh"; \
+        SHA256SUM="00938c3534750a0e4069499baf8f4e6dc1c2e471c86a59caa0dd03f4a9269db6"; \
+    elif [ "${UNAME_M}" = "s390x" ]; then \
+        MINICONDA_URL="https://repo.anaconda.com/miniconda/Miniconda3-${CONDA_VERSION}-Linux-s390x.sh"; \
+        SHA256SUM="a150511e7fd19d07b770f278fb5dd2df4bc24a8f55f06d6274774f209a36c766"; \
+    elif [ "${UNAME_M}" = "aarch64" ]; then \
+        MINICONDA_URL="https://repo.anaconda.com/miniconda/Miniconda3-${CONDA_VERSION}-Linux-aarch64.sh"; \
+        SHA256SUM="48a96df9ff56f7421b6dd7f9f71d548023847ba918c3826059918c08326c2017"; \
+    elif [ "${UNAME_M}" = "ppc64le" ]; then \
+        MINICONDA_URL="https://repo.anaconda.com/miniconda/Miniconda3-${CONDA_VERSION}-Linux-ppc64le.sh"; \
+        SHA256SUM="4c86c3383bb27b44f7059336c3a46c34922df42824577b93eadecefbf7423836"; \
+    fi && \
+    wget "${MINICONDA_URL}" -O miniconda.sh -q && \
+    echo "${SHA256SUM} miniconda.sh" > shasum && \
+    if [ "${CONDA_VERSION}" != "latest" ]; then sha256sum --check --status shasum; fi && \
+    mkdir -p /opt && \
+    bash miniconda.sh -b -p /opt/conda && \
+    rm miniconda.sh shasum && \
+    ln -s /opt/conda/etc/profile.d/conda.sh /etc/profile.d/conda.sh && \
+    echo ". /opt/conda/etc/profile.d/conda.sh" >> ~/.bashrc && \
+    echo "conda activate base" >> ~/.bashrc && \
+    find /opt/conda/ -follow -type f -name '*.a' -delete && \
+    find /opt/conda/ -follow -type f -name '*.js.map' -delete && \
+    /opt/conda/bin/conda clean -afy
+
+RUN conda create -n envd python=3.9
+
+ENV ENVD_PREFIX=/opt/conda/envs/envd/bin
+
+RUN update-alternatives --install /usr/bin/python python ${ENVD_PREFIX}/python 1 && \
+    update-alternatives --install /usr/bin/python3 python3 ${ENVD_PREFIX}/python3 1 && \
+    update-alternatives --install /usr/bin/pip pip ${ENVD_PREFIX}/pip 1 && \
+    update-alternatives --install /usr/bin/pip3 pip3 ${ENVD_PREFIX}/pip3 1
+
+COPY requirements.txt /
+
+RUN pip install -r requirements.txt
+
+RUN mkdir -p /workspace
+
+COPY . workspace/
+WORKDIR /workspace
+RUN pip install -e .[gpu]
+
+RUN modelz-llm --dry-run --model bigscience/bloomz-560m
+
+# disable huggingface update check (could be very slow)
+ENV HF_HUB_OFFLINE=true
+
+ENTRYPOINT [ "modelz-llm" ]
+CMD [ "--port", "8080" ]
diff --git a/images/base/Dockerfile → images/chatglm-6b-int4/Dockerfile b/images/base/Dockerfile → images/chatglm-6b-int4/Dockerfile
@@ -65,7 +65,7 @@ COPY . workspace/
 WORKDIR /workspace
 RUN pip install -e .[gpu]
 
-RUN modelz-llm --dry-run
+RUN modelz-llm --dry-run --model THUDM/chatglm-6b-int4
 
 # disable huggingface update check (could be very slow)
 ENV HF_HUB_OFFLINE=true

diff --git a/images/fastchat-t5-3b/Dockerfile b/images/fastchat-t5-3b/Dockerfile
@@ -65,10 +65,10 @@ COPY . workspace/
 WORKDIR /workspace
 RUN pip install -e .[gpu]
 
-RUN modelz-llm --dry-run --model lmsys/fastchat-t5-3b-v1.0
+# RUN modelz-llm --dry-run --model lmsys/fastchat-t5-3b-v1.0
 
-# disable huggingface update check (could be very slow)
-ENV HF_HUB_OFFLINE=true
+# # disable huggingface update check (could be very slow)
+# ENV HF_HUB_OFFLINE=true
 
 ENTRYPOINT [ "modelz-llm" ]
 CMD [ "--port", "8080", "--model", "lmsys/fastchat-t5-3b-v1.0" ]