Pre-trained weights

Can’t run Loading the pre-trained model in the week two assignment. I waited for time but it didn’t work

Get the model architecture

model = TransformerLM(mode=‘eval’)

Load the pre-trained weights

model.init_from_file(‘model.pkl.gz’, weights_only=True)

1 Like

Hi @Ahmad_khatatneh, @ebato,

Indeed, that is true. I have taken note of this and informed the team about it. It shall be fixed soon. I’ll inform you when it is. In case, I don’t please feel free to reach out to check it with me.

Thanks,
Mubsi

1 Like

Hi Mubsi,

I have the same issue, it still has this problem.

1 Like

Hi @zvika_sinkevich,

Can you share your lab ID with me ? In the assignment, when you click the top right “Help” button, a panel will open and your lab ID will be shown at the bottom.

I shall take a look.

When you reply back, kindly tag me in the post so that I’m notified.

Thanks,
Mubsi

@Mubsi

My lab Id: rnhsmklz
Thanks!

Hi @zvika_sinkevich,

Can you DM me your email ID which you use for Coursera ?

Thanks,
Mubsi

@Mubsi

I’ve sent you

Hi @Mubsi,

I though I’m having the same issue at first but then after some time it actually worked out. I think it would be great to add some note with estimate on how long it can take (Just like above for training).

Thanks,
Pavel

Hi @pavelkuchin,

How long do you think it took for you to get the pre-trained model ?

Thanks,
Mubsi

Hi @Mubsi,

I’ve just reloaded notebook and timed it. It takes about 14 minutes to get the model.

Thanks,
Pavel

Hi @pavelkuchin,

Hmm, thanks for letting me know. I thought it should take about 15 seconds to load it. I’ll inform this to the team.

Cheers,
Mubsi

I’m having the same issue with steps that include the pre-trained model. My notebook gets past loading the model in about 15 minutes, then it gets similarly stuck running “next_symbol”. My lab id is tbethntl.

Thanks

Hi, I encountered the same problem as well.
Has it been resolves?

No. But, I tried a little experiment. I submitted the notebook for grading and, voila!, the grade posted rather quickly. Assuming the grader ran the notebook the same way I had been doing, the problem appears to lie in my student environment, on jupyter.

I am also having this issue… Loading the pre-trained model hangs up indefinitely (I gave up after 15 mins… after reading this this thread)

Looks like there has been no response to previous posters for at least 5 days… Can somebody please look into this?

Thanks

UPDATE: For what it’s worth, I followed @Roger_Krimstock’s recommendation above, and just stopped the model-loading cell, and was still able to finish the notebook.

Same issue today on Jan-19. Finished after about 15 minutes. Then the next function wasn’t finishing after several minutes so I interrupted the kernel and tried again. Still slow, but finished the assignment.

Hi,

I have made a note asking people at the backend to resolve this issue. Hopefully it will be picked up soon.

1 Like

Same here. Notebook stuck at next_symbol

For anyone running into this issue, the model currently takes around 15 minutes to load. A nice chance to have a refreshing break :slight_smile: .

Hi all,

This behaviour is unpredictable. In my case, once it took 15 or so minutes. Next time, it took less than 10 seconds.

I’m not sure how to reproduce it, but once I was loading it, after a couple of minutes I restarted the kernel and tried again, but this time it loaded in mere seconds. Maybe you could also try this.

But yes, keep in mind that it sometimes might take several minutes.

We have tried to get to the bottom of this from our end, as well as asked for Coursera’s help, but since this is unpredictable, there wasn’t anything in particular we could put a finger on.

As @reinoudbosch mentioned, make the most of this time and take a break till it loads.

Best,
Mubsi