human-analysis / pytorchnet Goto Github PK

For "Adam" optimization method, "betas = (.., ..)" can not be set when I try to use "args.txt"
The error I am getting is:
ValueError: too many values to unpack (expected 2)

if someone wants to use this, a walk through example would be good

What is the expected structure of ImageNet dataset? I'm trying to pass train folder with folder per class train jpg images, and val folder with all 50k validation images. But it's not clear how to pass labels for val examples using datasets.ImageFolder in dataloader.py.

You have a serious problem with the way you report test accuracy. This caused incorrect results published in your PNN paper. See https://github.com/michaelklachko/pnn.pytorch for details.

I'm getting the following error when trying to run python main.py (using pytorch 4.1 and visdom 0.1.8.5):

Traceback (most recent call last):
  File "/home/michael/miniconda2/envs/pt/lib/python3.6/site-packages/visdom/__init__.py", line 446, in _send
    data=json.dumps(msg),
  File "/home/michael/miniconda2/envs/pt/lib/python3.6/json/__init__.py", line 231, in dumps
    return _default_encoder.encode(obj)
  File "/home/michael/miniconda2/envs/pt/lib/python3.6/json/encoder.py", line 199, in encode
    chunks = self.iterencode(o, _one_shot=True)
  File "/home/michael/miniconda2/envs/pt/lib/python3.6/json/encoder.py", line 257, in iterencode
    return _iterencode(o, 0)
  File "/home/michael/miniconda2/envs/pt/lib/python3.6/json/encoder.py", line 180, in default
    o.__class__.__name__)
TypeError: Object of type 'Scatter' is not JSON serializable