Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fire issue #2

Open
wyhinton opened this issue Mar 12, 2023 · 1 comment
Open

Fire issue #2

wyhinton opened this issue Mar 12, 2023 · 1 comment
Assignees
Labels
bug Something isn't working enhancement New feature or request good first issue Good for newcomers

Comments

@wyhinton
Copy link

wyhinton commented Mar 12, 2023

Hello. Cool project. Finally was able to get the right torch versions installed, and tryed to run, but Im getting an issue I think related to the cli?
Running python example_texture_nca.py --train_steps=500 run --lr=0.0005 as per the docs gives error:

GPU available: True, used: True
TPU available: False, using: 0 TPU cores
IPU available: False, using: 0 IPUs
Traceback (most recent call last):
  File "example_texture_nca.py", line 250, in <module>
    fire.Fire(FireRunner, name='runner')
  File "C:\Users\Primary User\anaconda3\envs\na3\lib\site-packages\fire\core.py", line 141, in Fire
    component_trace = _Fire(component, args, parsed_flag_args, context, name)
  File "C:\Users\Primary User\anaconda3\envs\na3\lib\site-packages\fire\core.py", line 466, in _Fire
    component, remaining_args = _CallAndUpdateTrace(
  File "C:\Users\Primary User\anaconda3\envs\na3\lib\site-packages\fire\core.py", line 681, in _CallAndUpdateTrace
    component = fn(*varargs, **kwargs)
TypeError: run() takes 1 positional argument but 18 were given```

any thoughts?
python 3.8.16

For a quick workaround I just did this:

  f = FireRunner()
  f.run()

Additional Note:
looks like the tensors are of type torch.int32. I needed to change them using .long().

        with torch.no_grad():
            self._organism_pool[idxs.long()] = y.to('cpu' if self.hparams.pool_on_cpu else self.device)
        with torch.no_grad():
            # reset runs in the pool
            if i % self.hparams.pool_reset_element_period == 0:
                self._organism_pool[idxs[0]] = self.nca.make_start_batch(batch_size=1, size=self.hparams.nca_img_size, device=self._organism_pool.device)[0]
            # load incomplete runs from the pool
   
            x = self._organism_pool[idxs.long()].to(self.device)
            x_img = self.nca.extract_rgb(x)
@nmichlo
Copy link
Owner

nmichlo commented Mar 12, 2023

Thank you for investigating the errors!

Versions of all the libraries are quite out of date. Will open a PR soon to update the versions. Your feedback will definitely help with this!

If things are still not working, would try and roll back versions to around the time the last commits happened.

@nmichlo nmichlo added bug Something isn't working enhancement New feature or request good first issue Good for newcomers labels Mar 13, 2023
@nmichlo nmichlo self-assigned this Oct 13, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working enhancement New feature or request good first issue Good for newcomers
Projects
None yet
Development

No branches or pull requests

2 participants