Update habana docker image version and DeepSpeed dependency to 1.21.0 (#2213)

Signed-off-by: Sun, Xuehao <[email protected]>
Co-authored-by: Xin He <[email protected]>

Author: XuehaoSun

.azure-pipelines/scripts/ut/3x/run_3x_pt_fp8.sh

@@ -12,6 +12,7 @@ echo "##[section]import check pass"
 # install requirements
 echo "##[group]set up UT env..."
 export LD_LIBRARY_PATH=/usr/local/lib/:$LD_LIBRARY_PATH
+export PT_HPU_LAZY_MODE=1
 sed -i '/^intel_extension_for_pytorch/d' /neural-compressor/test/3x/torch/requirements.txt
 sed -i '/^auto_round/d' /neural-compressor/test/3x/torch/requirements.txt
 cat /neural-compressor/test/3x/torch/requirements.txt

.azure-pipelines/template/docker-template.yml

@@ -74,7 +74,7 @@ steps:
 
   - ${{ if eq(parameters.imageSource, 'pull') }}:
       - script: |
-          docker pull vault.habana.ai/gaudi-docker/1.20.0/ubuntu22.04/habanalabs/pytorch-installer-2.6.0:latest
+          docker pull vault.habana.ai/gaudi-docker/1.21.0/ubuntu22.04/habanalabs/pytorch-installer-2.6.0:latest
         displayName: "Pull habana docker image"
 
   - script: |
@@ -95,7 +95,7 @@ steps:
             else
                 docker run -dit --disable-content-trust --privileged --name=${{ parameters.containerName }} --shm-size="2g" \
                 --runtime=habana -e HABANA_VISIBLE_DEVICES=all -e OMPI_MCA_btl_vader_single_copy_mechanism=none --cap-add=sys_nice --net=host --ipc=host \
-                -v ${BUILD_SOURCESDIRECTORY}:/neural-compressor vault.habana.ai/gaudi-docker/1.20.0/ubuntu22.04/habanalabs/pytorch-installer-2.6.0:latest
+                -v ${BUILD_SOURCESDIRECTORY}:/neural-compressor vault.habana.ai/gaudi-docker/1.21.0/ubuntu22.04/habanalabs/pytorch-installer-2.6.0:latest
                 docker exec ${{ parameters.containerName }} bash -c "ln -sf \$(which python3) /usr/bin/python"
             fi
             echo "Show the container list after docker run ... "

README.md

@@ -56,8 +56,11 @@ To try on Intel Gaudi2, docker image with Gaudi Software Stack is recommended, p
 
 Run a container with an interactive shell, [more info](https://docs.habana.ai/en/latest/Installation_Guide/Additional_Installation/Docker_Installation.html#docker-installation)
 ```
-docker run -it --runtime=habana -e HABANA_VISIBLE_DEVICES=all -e OMPI_MCA_btl_vader_single_copy_mechanism=none --cap-add=sys_nice --net=host --ipc=host vault.habana.ai/gaudi-docker/1.20.0/ubuntu24.04/habanalabs/pytorch-installer-2.6.0:latest
+docker run -it --runtime=habana -e HABANA_VISIBLE_DEVICES=all -e OMPI_MCA_btl_vader_single_copy_mechanism=none --cap-add=sys_nice --net=host --ipc=host vault.habana.ai/gaudi-docker/1.21.0/ubuntu24.04/habanalabs/pytorch-installer-2.6.0:latest
 ```
+
+> Note: Since Habana software >= 1.21.0, `PT_HPU_LAZY_MODE=0` is the default setting. However, most low-precision functions (such as `convert_from_uint4`) do not support this setting. Therefore, we recommend setting `PT_HPU_LAZY_MODE=1` to maintain compatibility.
+
 Run the example,
 ```python
 from neural_compressor.torch.quantization import (

test/3x/torch/requirements.txt

@@ -1,6 +1,6 @@
 auto_round
 datasets
-deepspeed @ git+https://github.com/HabanaAI/DeepSpeed.git@1.20.0
+deepspeed @ git+https://github.com/HabanaAI/DeepSpeed.git@1.21.0
 expecttest
 intel_extension_for_pytorch
 numpy