Hi,
I tried running the Decoder only model locally but I am having the ‘the’ problem-- Or basically my model only produces (repeatingly) either ‘First Citizen’ or occasionally ‘Second Citizen’ and nothing more.
My training run appears to be working okay.
I know running the scripts on your own machine is ‘buyer beware’, but I didn’t know if anyone had any idea what the problem might be?
is this lab codes from pytorch course you selected??
in the section where first citizen is seen, look at the problem column, what does it show?
did you download all the files?
I think it was my Shakespeare file was for some reason very truncated (!). It is showing better peram’s now, but is going very slow on my mac.
I’d post the codes here (sans the rest of the notebook-- and it is not an assignment, all codes are provided)-- But I don’t know if that is against the rules.
is it related to the course category you have selected @Nevermnd ??
if not, kindly move it ai discussion category.
more than sharing codes, actually if you can provide information on from where are the codes, how are you running the codes in console terminal. Did ihe code files have any requirement ttxt.file ?
Sorry, asking these questions because it is confusing as to why the decoder file codes in pytorch course 3 and now you are saying you are getting better output but slow, so probably it is more related to your GPU capacity?
why didn’t you first try running on your Google colab, that would have solved your gpu usage issue. Just trying to understand your issue. kindly provide more brief explanation of what you are trying to do, what all steps you followed, what you wanted but what you got and what you want to get.
regards
Dr. Deepti
@Deepti_Prasad I do have an RTX equipped desktop. I am trying to run this on my M1 Mac Book. It is okay, I think I solved the problem. I think I am ‘okay’ posting these as there is no student written code, but this is everything outside the ‘text’.
Decoder.py (13.1 KB)
helper_utils.py (20.5 KB)
first confirm is this issue related to pytorch? course
and if you have solved your issue, then there is nothing to ponder further right.
