pytorch-ts
pytorch-ts copied to clipboard
zalandoresearch
TypeError: __init__() got an unexpected keyword argument 'freq'
Hi, Training is well when I reproduce the tests, but in forecasting, I encountered an issue. Here are some specifics. (the version of packages: pytorchts-0.6.0, gluonts-0.10.0)
/home/axx/anaconda3/envs/torch2/lib/python3.8/site-packages/gluonts/dataset/multivariate_grouper.py:191: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.
return {FieldName.TARGET: np.array([funcs(data) for data in dataset])}
100%|███████████████████████████████████████████████████████| 99/99 [15:13<00:00, 9.23s/it, epoch=1/1, avg_loss=-64.6]
Traceback (most recent call last):
File "Multivariate-Flow-Solar.py", line 202, in
When everyone faces the same problem, any assistance is greatly appreciated.
Can you kindly downgrade gluonts as I have not updated this lib. to use the new APIs etc.
@kashif Thank you very much for your generous answers. Downgrading gluons from 0.10.0 to 0.9 works for me. By the way, if you downgrade your gluonts to 0.9, don't forget to change the file of ditribution_out.py in line 34(i.e. gluonts.torch.distributions.distribution_output-->gluonts.torch.modules.distribution_output)
Hi, @kashif, I sent you an email late last night, and I would appreciate it if you could find some time in your hectic work schedule to respond. Your study is really great, and I'll be basing my own scientific research on it!
Hi, @kashif Does the implementation of this code not use the multi-scale structure like the one in the Real-NVP paper?
@hanlaoshi not sure if i understand, but what do you mean by " multi-scale structure" exactly? thanks!
