Error : Training the compressed model in my own local server.

[kbjsniper]

Hi, Thank you for your great works about Deep learning.
I've explored the NetsPresso through the web.
I've tested the module of compression, after I compressed the model, downloaded that model, and I got the error.
In detail, I compressed the model with ratio 0.5 which the original model was sample_yolov5s_voc that you provide in your web,
I downloaded that model, when I tried to train the compressed model, but I got the error as below.

Traceback (most recent call last):
File "train.py", line 642, in
main(opt)
File "train.py", line 531, in main
train(opt.hyp, opt, device, callbacks)
File "train.py", line 127, in train
csd = ckpt['model'].float().state_dict() # checkpoint state_dict as FP32
TypeError: 'GraphModule' object is not subscriptable
Traceback (most recent call last):
File "train.py", line 642, in
main(opt)
File "train.py", line 531, in main
train(opt.hyp, opt, device, callbacks)
File "train.py", line 127, in train
csd = ckpt['model'].float().state_dict() # checkpoint state_dict as FP32
TypeError: 'GraphModule' object is not subscriptable
YOLOv5s summary: 214 layers, 7105933 parameters, 7105933 gradients, 16.2 GFLOPs

Traceback (most recent call last):
File "train.py", line 642, in
main(opt)
File "train.py", line 531, in main
train(opt.hyp, opt, device, callbacks)
File "train.py", line 127, in train
csd = ckpt['model'].float().state_dict() # checkpoint state_dict as FP32
TypeError: 'GraphModule' object is not subscriptable
Traceback (most recent call last):
File "train.py", line 642, in
main(opt)
File "train.py", line 531, in main
train(opt.hyp, opt, device, callbacks)
File "train.py", line 127, in train
csd = ckpt['model'].float().state_dict() # checkpoint state_dict as FP32
TypeError: 'GraphModule' object is not subscriptable
ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 17424) of binary: /DATA/bjkang/tfl_py375/bin/python
Traceback (most recent call last):
File "/usr/lib/python3.7/runpy.py", line 193, in _run_module_as_main
"main", mod_spec)
File "/usr/lib/python3.7/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/DATA/bjkang/tfl_py375/lib/python3.7/site-packages/torch/distributed/launch.py", line 195, in
main()
File "/DATA/bjkang/tfl_py375/lib/python3.7/site-packages/torch/distributed/launch.py", line 191, in main
launch(args)
File "/DATA/bjkang/tfl_py375/lib/python3.7/site-packages/torch/distributed/launch.py", line 176, in launch
run(args)
File "/DATA/bjkang/tfl_py375/lib/python3.7/site-packages/torch/distributed/run.py", line 756, in run
)(*cmd_args)
File "/DATA/bjkang/tfl_py375/lib/python3.7/site-packages/torch/distributed/launcher/api.py", line 132, in call
return launch_agent(self._config, self._entrypoint, list(args))
File "/DATA/bjkang/tfl_py375/lib/python3.7/site-packages/torch/distributed/launcher/api.py", line 248, in launch_agent
failures=result.failures,
torch.distributed.elastic.multiprocessing.errors.ChildFailedError:

Please, help me about this problem.
Thank in advance.

[daechanim]

Hi kbjsniper,

Sorry for the late reply.
Thank you for using NetsPresso.

NetsPresso offers a feature to retrain models after compression, making it easy for you to conduct retraining effortlessly.
You can access the retraining functionality directly through the retrain button on the list of 'Models'.

[kbjsniper]

Hi, daechanim

Thank you for your reply.

I'd like to train the compressed model in my own server after downloading from your Web, because I have custom datasets that are very big size.
Could you help me?

[daechanim]

We had a meeting regarding this, but I'm sorry for not reporting it here.