Transformers in requirements.txt but Error -> No module named 'transformers'

[Charly_Wargnier]

Hi guys,

I’ve added “Transformers” in the requirements.txt file, but I got a ModuleNotFoundError -> No module named 'transformers' when I’m trying to deploy in Streamlit share.

Not sure what I’m doing wrong here. Here’s the app if you wanted to have a glance:

https://share.streamlit.io/charlywargnier/gpuoneshottransformersstreamlit/main/app.py

Thanks,
Charly

[randyzwitch]

Check the sidebar for an error message. We launch the app even if a pip failure occurs, so there’s probably a message that some driver was missing or something else to cause the requirements not to get installed properly.

[Charly_Wargnier]

I see, thanks Randy!

So I’ve checked and removed the errors seen in the log file. No more ModuleNotFoundError -> No module named 'transformers' - great news!

The app shows another issue, however, and the installation seems to keep going/never stops:

I pasted the log here: https://pastebin.com/HzEcvN1k

I can’t seem to find any error here, except:

[manager] Error checking Streamlit healthz: Get "http://localhost:8501/healthz": read tcp 127.0.0.1:48364->127.0.0.1:8501: read: connection reset by peer

I’ll keep trying various things over the weekend!

Thanks,
Charly

[Charly_Wargnier]

Quick heads-up @randyzwitch

15 tries later, it’s still not working and the installation keeps looping it seems …
(Latest error log is here, as well as the repo)

I think I know where the issue is coming from yet not sure how to fix it.

When I try the same script in Google Colab, Colab downloads all the needed weights/models.

I believe we need a way to upload these models to Streamlit Sharing, yet not sure how to do it:

Maybe via an extra text file?

Thanks,
Charly

[randyzwitch]

Do you have a feeling for the cumulative size of the models? I’m wondering if you’re exceeding the VM size, and it’s caught in an unfortunate deploy cycle

[Charly_Wargnier]

It relies on GPU PyTorch. To give you an idea of size, here’s the error I get in the free version of Heroku (capped at 500M):

Compiled slug size: 863.7M is too large (max is 500M).

Thanks,
Charly

[randyzwitch]

Currently, we’re at 800MB for the limits

Deploy an app — Streamlit 0.69.2 documentation

[Charly_Wargnier]

Thanks Randy.

So I’ve reduced size to 247MB (according to Heroku) by downloading CPU only wheels:

It works fine in Heroku yet still the same looping issue in Streamlit Sharing…

[manuel_illanes]

How did you solve this importing error? I’m getting the same in my local computer, I’ve reinstalled streamlit 3 times now

[someshfengde]

did anyone found solution for this ?

[snehankekre]

Hi @someshfengde

Could you please describe your issue in more detail and perhaps share a screenshot showing the error you’re running into? That will help us reproduce the error. Have you checked if there’s a typo in your requirements.txt file?

As an example, I created a sample app using the transformers module and successfully deployed it on sharing:

snehankekre/transformers-test

https://share.streamlit.io/snehankekre/transformers-test/main/transformers_sharing.py

Best,
Snehan

[someshfengde]

Hey did this but it gave me size limit exceeded error ( I understand that cause my model was like 500 mb)

you can check the code here

github repo : GitHub - someshfengde/Signalism_project: building the streamlit application for generating the text output while the input is provided in the form of textual data

[Charly_Wargnier]

Hey Snehan!

I got an error on your app when tried this morning:
https://share.streamlit.io/snehankekre/transformers-test/main/transformers_sharing.py

Is this expected?

Thanks
Charly

[snehankekre]

Thanks for flagging, @Charly_Wargnier

I tested it locally but didn’t do so on sharing. I fixed the error by including tensorflow in my requirements.txt file.

It should work now:

Best,
Snehan

[snehankekre]

Hi @someshfengde,

Apps deployed on Streamlit sharing do have some resource limits. One of them is 1 GB of RAM per app. I suspect the GPT2 model takes up much more than 1 GB of RAM when read into memory, leading to the error.

Best,
Snehan

[someshfengde]

ohh thanks for the reply I’ll try to build the app with some smaller model

[Charly_Wargnier]

It does! Thank you!

[Charly_Wargnier]

Thanks Snehan!

JFYI, I’ve just tried and it’s now a 404

https://share.streamlit.io/snehankekre/transformers-test/main/transformers_sharing.py

[snehankekre]

@Charly_Wargnier That’s expected I created the app temporarily to demonstrate a fix for the No module named 'transformers' error. I want to avoid a demo eating up sharing resources.

[Charly_Wargnier]

Ok, cheers