How to add a new model ?

[Coderx7]

Hi, I tried adding a new model to no avail.
what I did is as follows,

1. added the model name to torchdistill/models/classification/__init__.py
2. added from ..registry import register_model_func_func to the model definition (here simplenet)
3. created the required methods and annotated it with @register_model_func like this:

@register_model_func
def smpnetv1_05m_m1(pretrained_cfg: str = 'smpnetv1_05m_m1', pretrained: bool = False, **kwargs: Any) -> SimpleNet:
    model_args = dict(**kwargs)
    return _create_simpnet(1000, scale=0.5, network_idx=17, mode=1, pretrained=pretrained)

and saved everything.
now when trying to run an example I get this error message:

ImportError: cannot import name 'register_model_func_func' from 'torchdistill.models.registry' (/home/hossein/torchdistill/torchdistill/models/registry.py)

here is the full error message (I was trying to load an existing example before working with the new model, but nevertheless, the error pops up regardless)

(base) hossein@hossein-pc:~/torchdistill$ python -m torch.distributed.launch --nproc_per_node=1 --use_env examples/image_classification.py --world_size 1 --config configs/sample/ilsvrc2012/single_stage/kd/alexnet_from_resnet152.yaml --log log/ilsvrc2012/kd/alexnet_from_resnet152.txt
/home/hossein/anaconda3/lib/python3.9/site-packages/torch/distributed/launch.py:178: FutureWarning: The module torch.distributed.launch is deprecated
and will be removed in future. Use torchrun.
Note that --use_env is set by default in torchrun.
If your script expects `--local_rank` argument to be set, please
change it to read from `os.environ['LOCAL_RANK']` instead. See 
https://pytorch.org/docs/stable/distributed.html#launch-utility for 
further instructions

  warnings.warn(
Traceback (most recent call last):
  File "/home/hossein/torchdistill/examples/image_classification.py", line 15, in <module>
    from torchdistill.core.distillation import get_distillation_box
  File "/home/hossein/torchdistill/torchdistill/core/__init__.py", line 1, in <module>
    from . import pre_epoch_proc, pre_forward_proc, forward_proc, post_forward_proc, post_epoch_proc
  File "/home/hossein/torchdistill/torchdistill/core/post_epoch_proc.py", line 6, in <module>
    from ..models.wrapper import AuxiliaryModelWrapper
  File "/home/hossein/torchdistill/torchdistill/models/__init__.py", line 2, in <module>
    from .classification import CLASSIFICATION_MODEL_FUNC_DICT
  File "/home/hossein/torchdistill/torchdistill/models/classification/__init__.py", line 1, in <module>
    from . import densenet, resnet, wide_resnet,simplenet
  File "/home/hossein/torchdistill/torchdistill/models/classification/simplenet.py", line 10, in <module>
    from ..registry import register_model_func_func
ImportError: cannot import name 'register_model_func_func' from 'torchdistill.models.registry' (/home/hossein/torchdistill/torchdistill/models/registry.py)
ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 372873) of binary: /home/hossein/anaconda3/bin/python
Traceback (most recent call last):
  File "/home/hossein/anaconda3/lib/python3.9/runpy.py", line 197, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/home/hossein/anaconda3/lib/python3.9/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/home/hossein/anaconda3/lib/python3.9/site-packages/torch/distributed/launch.py", line 193, in <module>
    main()
  File "/home/hossein/anaconda3/lib/python3.9/site-packages/torch/distributed/launch.py", line 189, in main
    launch(args)
  File "/home/hossein/anaconda3/lib/python3.9/site-packages/torch/distributed/launch.py", line 174, in launch
    run(args)
  File "/home/hossein/anaconda3/lib/python3.9/site-packages/torch/distributed/run.py", line 715, in run
    elastic_launch(
  File "/home/hossein/anaconda3/lib/python3.9/site-packages/torch/distributed/launcher/api.py", line 131, in __call__
    return launch_agent(self._config, self._entrypoint, list(args))
  File "/home/hossein/anaconda3/lib/python3.9/site-packages/torch/distributed/launcher/api.py", line 245, in launch_agent
    raise ChildFailedError(
torch.distributed.elastic.multiprocessing.errors.ChildFailedError: 
============================================================
examples/image_classification.py FAILED
------------------------------------------------------------
Failures:
  <NO_OTHER_FAILURES>
------------------------------------------------------------
Root Cause (first observed failure):
[0]:
  time      : 2023-01-16_19:28:07
  host      : hossein-pc
  rank      : 0 (local_rank: 0)
  exitcode  : 1 (pid: 372873)
  error_file: <N/A>
  traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
============================================================

How should I go about this?
Thanks a lot in advance

[yoshitomo-matsubara]

Hi @Coderx7 ,

It seems you edited files under torchdistill/ in this repo. Note that torchdistill is available as a PyPI package and installable (pip install torchdistill), and I'm currently working on next release and editing files under torchdistill/.
For this reason, I suggest you do not edit torchdistill/ to add new modules, but instead it is designed to add new modules by registry function like you did (almost there, but you edited torchdistill/ files)

1. Uninstall torchdistill pip uninstall torchdistill if you installed the local torchdistill instead of one available at PyPI
2. Reinstall torchdistill via PyPI pip install torchdistill
3. Create a Python file somewhere under examples/ (say examples/my_module.py) that contains the following

from torchdistill.models.registry import register_model_func

# Some import statements here to call _create_simpnet

@register_model_func
def smpnetv1_05m_m1(pretrained_cfg: str = 'smpnetv1_05m_m1', pretrained: bool = False, **kwargs: Any) -> SimpleNet:
    model_args = dict(**kwargs)
    return _create_simpnet(1000, scale=0.5, network_idx=17, mode=1, pretrained=pretrained)

4. Have examples/image_classification.py import your new Python code at somewhere around here
   e.g., if your new Python file is saved as examples/my_module.py,

import my_module

6. Run examples/image_classification.py

[lansiyuan3]

HI@Coderx7
I mentioned your question，and i want to do something like you .I would like to ask you the exact steps to create a new model.
Thank you!