size mismatch for emb.weight: copying a param with shape torch.Size([25, 20]) from checkpoint, the shape in current model is torch.Size([21, 20]).

[marcos914]

When I use the prediction_emb.py to predict rt value ,I have found some questions.How can I deal with it.
python prediction_emb.py 46 param/dia_all_epo20_dim24_conv12/dia_all_epo20_dim24_conv12_filled.pt 12 data/SCX.txt

Traceback (most recent call last):
File "prediction_emb.py", line 78, in
obse,pred1=pred_from_model(conv1,conv1,round1model,RTtest,15)
File "prediction_emb.py", line 17, in pred_from_model
model.load_state_dict(torch.load(param_path))
File "/home/renzhe/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 769, in load_state_dict
self.class.name, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for CapsuleNet:
size mismatch for emb.weight: copying a param with shape torch.Size([25, 20]) from checkpoint, the shape in current model is torch.Size([21, 20]).

[horsepurve]

Could you try changing config.py from

dict_path = ''
max_length = 50

to

dict_path = 'data/mod.txt'
max_length = 66

because 66 is the max length in the "dia.txt" dataset, while 50 is for the "mod.txt" dataset, and "dict_path" means we want to include four modifications.

By the way, I guess you were using RPLC model to make predictions on SCX data, which might not be valid, since RP and SCX are different separation mechanisms...

[nasseralfadly]

I have the same problem in my model