Alibaba Cloud Linux:Use Alibaba Cloud Linux 3 AI-enhanced images

1. Enable the keentune optimization tool.

   The keentune optimization tool is pre-installed on the Alibaba Cloud Linux 3 AI Extension Edition image. This tool provides optimizations for various scenarios. Follow these steps to enable optimization for AI scenarios.

   systemctl stop tuned
   systemctl disable tuned
   systemctl start keentune-target
   systemctl enable keentune-target
   systemctl enable keentuned
   systemctl start keentuned
   keentune profile set ai_train.profile

The keentune optimizations require an OS restart to take effect. To disable the optimizations, run the keentune profile rollback command. This change also requires an OS restart.

2. Install Docker.

   To install Docker and its related components, see Train models using PyTorch GPU images.

3. Obtain the test image.

   docker pull ac2-registry.cn-hangzhou.cr.aliyuncs.com/ac2/openclip-bevformer:v0.1-torch2.6-cuda12.6-py3.10-ubuntu22.04

4. Download the datasets.

   The container image does not include datasets. After you download the image, run the following commands to download the required model datasets:

   OpenCLIP training and inference datasets

   # Download the training dataset
   mkdir -p /workspace/dataset && cd /workspace/dataset
   wget -O mscoco.parquet https://hf-mirror.com/datasets/ChristophSchuhmann/MS_COCO_2017_URL_TEXT/resolve/main/mscoco.parquet?download=true
   pip3 install img2dataset webdataset==0.2.86 numpy==1.23.5 --ignore-installed && NO_ALBUMENTATIONS_UPDATE=1 img2dataset --url_list mscoco.parquet --input_format "parquet" --url_col "URL" --caption_col "TEXT" --output_format webdataset --output_folder COCO_2017_Captions-webdataset-592k-256x256-296shards --processes_count 16 --thread_count 64 --image_size 256 --number_sample_per_shard 2000
   # Download the inference dataset
   mkdir -p /workspace/dataset && cd /workspace/dataset
   wget https://image-net.org/data/ILSVRC/2012/ILSVRC2012_img_val.tar --no-check-certificate
   mkdir -p ILSVRC2012_img_val
   tar xvf ILSVRC2012_img_val.tar -C ILSVRC2012_img_val
   cd ILSVRC2012_img_val/ && wget https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh && bash valprep.sh

   For more information, see GitHub - mlfoundations/open_clip: An open source implementation of CLIP.

   BEVFormer training dataset

   mkdir -p /workspace/BEVFormer/data && cd /workspace/BEVFormer/data
   wget https://d36yt3mvayqw5m.cloudfront.net/public/v1.0/v1.0-mini.tgz
   mkdir -p nuscenes && tar -xzf v1.0-mini.tgz -C ./nuscenes/
   wget https://d36yt3mvayqw5m.cloudfront.net/public/v1.0/can_bus.zip
   unzip -q can_bus.zip
   cd /workspace/BEVFormer && python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./data/nuscenes --extra-tag nuscenes --version v1.0-mini --canbus ./data

   For more information, see BEVFormer/docs/prepare_dataset.md at master · fundamentalvision/BEVFormer.

5. Run tests in the container image.

   OpenCLIP

   • Training command:

     cd /workspace/open_clip/src && torchrun --nproc_per_node 8 -m open_clip_train.main --model RN50 --train-data /workspace/dataset/COCO_2017_Captions-webdataset-592k-256x256-296shards/\{00000..00295\}.tar --train-num-samples 591753 --dataset-type webdataset --batch-size 1152 --precision amp --workers 8 --epochs 4 --log-every-n-steps 1 --torchcompile

   • Inference command:

     cd /workspace/open_clip/src && torchrun --nproc_per_node 1 -m open_clip_train.main --imagenet-val /workspace/dataset/ILSVRC2012_img_val --model RN50 --batch-size 1152 --workers 8 --pretrained openai

   BEVFormer

   Training command:

   cd /workspace/BEVFormer && TORCH_ALLOW_TF32_CUBLAS_OVERRIDE=1 tools/dist_train.sh projects/configs/bevformer/bevformer_base.py 8