Run the yesno recipe with docker in GitHub actions (#1191)

.github/workflows/run-docker-image.yml

@@ -25,12 +25,23 @@ jobs:
         uses: addnab/docker-run-action@v3
         with:
             image: k2fsa/icefall:${{ matrix.image }}
+            shell: bash
             run: |
               uname -a
               cat /etc/*release
 
               nvcc --version
 
+              # For torch1.9.0-cuda10.2
+              export LD_LIBRARY_PATH=/usr/local/cuda-10.2/compat:$LD_LIBRARY_PATH
+
+              # For torch1.12.1-cuda11.3
+              export LD_LIBRARY_PATH=/usr/local/cuda-11.3/compat:$LD_LIBRARY_PATH
+
+              # For torch2.0.0-cuda11.7
+              export LD_LIBRARY_PATH=/usr/local/cuda-11.7/compat:$LD_LIBRARY_PATH
+
+
               which nvcc
               cuda_dir=$(dirname $(which nvcc))
               echo "cuda_dir: $cuda_dir"
@@ -40,20 +51,26 @@ jobs:
 
               find / -name libcuda.so* 2>/dev/null
 
-              pushd /opt/conda/lib/stubs && ln -s libcuda.so libcuda.so.1 && popd
+              # for torch1.13.0-cuda11.6
+              if [ -e /opt/conda/lib/stubs/libcuda.so ]; then
+                cd /opt/conda/lib/stubs && ln -s libcuda.so libcuda.so.1 && cd -
+                export LD_LIBRARY_PATH=/opt/conda/lib/stubs:$LD_LIBRARY_PATH
+              fi
 
-              export LD_LIBRARY_PATH=/opt/conda/lib/stubs:$LD_LIBRARY_PATH
+              find / -name libcuda.so* 2>/dev/null
-              echo "LD_LIBRARY_PATH $LD_LIBRARY_PATH"
+              echo "LD_LIBRARY_PATH: $LD_LIBRARY_PATH"
 
               python3 --version
               which python3
 
+              python3 -m pip list
+
               echo "----------torch----------"
               python3 -m torch.utils.collect_env
 
               echo "----------k2----------"
               python3 -c "import k2; print(k2.__file__)"
-              python3 -c "import k2; print(k2.__version__)"
+              python3 -c "import k2; print(k2.__dev_version__)"
               python3 -m k2.version
 
               echo "----------lhotse----------"
@@ -64,3 +81,12 @@ jobs:
               python3 -c "import kaldifeat; print(kaldifeat.__file__)"
               python3 -c "import kaldifeat; print(kaldifeat.__version__)"
 
+              echo "Test yesno recipe"
+
+              cd egs/yesno/ASR
+
+              ./prepare.sh
+
+              ./tdnn/train.py
+
+              ./tdnn/decode.py

docker/torch1.12.1-cuda11.3.dockerfile

@@ -1,4 +1,4 @@
-FROM pytorch/pytorch:1.12.1-cuda11.3-cudnn8-runtime
+FROM pytorch/pytorch:1.12.1-cuda11.3-cudnn8-devel
 
 ENV LC_ALL C.UTF-8
 
@@ -51,7 +51,15 @@ RUN pip install --no-cache-dir \
       sentencepiece>=0.1.96 \
       tensorboard \
       typeguard \
-      dill
+      dill \
+      onnx \
+      onnxruntime \
+      onnxmltools \
+      multi_quantization \
+      typeguard \
+      numpy \
+      pytest \
+      graphviz
 
 RUN git clone https://github.com/k2-fsa/icefall /workspace/icefall && \
     cd /workspace/icefall && \

docker/torch1.13.0-cuda11.6.dockerfile

@@ -51,7 +51,15 @@ RUN pip install --no-cache-dir \
       sentencepiece>=0.1.96 \
       tensorboard \
       typeguard \
-      dill
+      dill \
+      onnx \
+      onnxruntime \
+      onnxmltools \
+      multi_quantization \
+      typeguard \
+      numpy \
+      pytest \
+      graphviz
 
 RUN git clone https://github.com/k2-fsa/icefall /workspace/icefall && \
     cd /workspace/icefall && \

docker/torch1.9.0-cuda10.2.dockerfile

@@ -1,4 +1,4 @@
-FROM pytorch/pytorch:1.9.0-cuda10.2-cudnn7-runtime
+FROM pytorch/pytorch:1.9.0-cuda10.2-cudnn7-devel
 
 ENV LC_ALL C.UTF-8
 
@@ -13,6 +13,13 @@ LABEL k2_version=${K2_VERSION}
 LABEL kaldifeat_version=${KALDIFEAT_VERSION}
 LABEL github_repo="https://github.com/k2-fsa/icefall"
 
+# see https://developer.nvidia.com/blog/updating-the-cuda-linux-gpg-repository-key/
+
+RUN rm /etc/apt/sources.list.d/cuda.list && \
+	rm /etc/apt/sources.list.d/nvidia-ml.list && \
+	apt-key del 7fa2af80
+
+
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
         curl \
@@ -37,8 +44,15 @@ RUN apt-get update && \
         zlib1g-dev \
         && rm -rf /var/lib/apt/lists/*
 
+RUN wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-keyring_1.0-1_all.deb && \
+    dpkg -i cuda-keyring_1.0-1_all.deb && \
+    rm -v cuda-keyring_1.0-1_all.deb && \
+    apt-get update && \
+    rm -rf /var/lib/apt/lists/*
+
 # Install dependencies
-RUN pip install --no-cache-dir \
+RUN pip uninstall -y tqdm && \
+    pip install -U --no-cache-dir \
       torchaudio==${TORCHAUDIO_VERSION} -f https://download.pytorch.org/whl/torch_stable.html \
       k2==${K2_VERSION} -f https://k2-fsa.github.io/k2/cuda.html \
       kaldifeat==${KALDIFEAT_VERSION} -f https://csukuangfj.github.io/kaldifeat/cuda.html \
@@ -51,7 +65,17 @@ RUN pip install --no-cache-dir \
       sentencepiece>=0.1.96 \
       tensorboard \
       typeguard \
-      dill
+      dill \
+      onnx \
+      onnxruntime \
+      onnxmltools \
+      multi_quantization \
+      typeguard \
+      numpy \
+      pytest \
+      graphviz \
+      tqdm>=4.63.0
+
 
 RUN git clone https://github.com/k2-fsa/icefall /workspace/icefall && \
     cd /workspace/icefall && \

docker/torch2.0.0-cuda11.7.dockerfile

@@ -1,4 +1,4 @@
-FROM pytorch/pytorch:2.0.0-cuda11.7-cudnn8-runtime
+FROM pytorch/pytorch:2.0.0-cuda11.7-cudnn8-devel
 
 ENV LC_ALL C.UTF-8
 
@@ -51,7 +51,15 @@ RUN pip install --no-cache-dir \
       sentencepiece>=0.1.96 \
       tensorboard \
       typeguard \
-      dill
+      dill \
+      onnx \
+      onnxruntime \
+      onnxmltools \
+      multi_quantization \
+      typeguard \
+      numpy \
+      pytest \
+      graphviz
 
 RUN git clone https://github.com/k2-fsa/icefall /workspace/icefall && \
     cd /workspace/icefall && \