Skip to main content

Overview

SSCMA currently supports the following models. You can refer to the corresponding tutorials to complete the training of the models and obtain the model weights.

tip

Before start training, we recommend you to read Config and Datasets sections first.

Parameter Descriptions​

For more parameters during model training, you can refer the code below.

python3 tools/train.py --help

# Train SSCMA models

# positional arguments:
#   config                the model config file path

# optional arguments:
#   -h, --help            show this help message and exit
#   --work_dir WORK_DIR, --work-dir WORK_DIR
#                         the directory to save logs and models
#   --amp                 enable automatic-mixed-precision during training (https://pytorch.org/tutorials/recipes/recipes/amp_recipe.html)
#   --auto_scale_lr, --auto-scale-lr
#                         enable automatic-scale-LR during training
#   --resume [RESUME]     resume training from the checkpoint of the last epoch (or a specified checkpoint path)
#   --no_validate, --no-validate
#                         disable checkpoint evaluation during training
#   --launcher {none,pytorch,slurm,mpi}
#                         the job launcher for MMEngine
#   --cfg_options CFG_OPTIONS [CFG_OPTIONS ...], --cfg-options CFG_OPTIONS [CFG_OPTIONS ...]
#                         override some settings in the used config, the key-value pair in 'xxx=yyy' format will be merged into config file
#   --local_rank LOCAL_RANK, --local-rank LOCAL_RANK
#                         set local-rank for PyTorch
#   --dynamo_cache_size DYNAMO_CACHE_SIZE, --dynamo-cache-size DYNAMO_CACHE_SIZE
#                         set dynamo-cache-size limit for PyTorch
#   --input_shape INPUT_SHAPE [INPUT_SHAPE ...], --input-shape INPUT_SHAPE [INPUT_SHAPE ...]
#                         Extension: input data shape for model parameters estimation, e.g. 1 3 224 224

Deployment​

After exporting the model, you can deploy the model to an edge computing device for testing and evaluation. You can refer to Deploy section to learn more about how to deploy the model.

Loading Comments...