🐛 Bug: Environment does not exist error during fetch for model eos85a3 on google collab

[paulinebanye]

Describe the bug.

Hi @GemmaTuron, I received this error while fetching model eos85a3 on collab Exception: eos85a3 environment does not exist. Link to eos85a3 ipynb file

I believe the error is from the model because I was able to fetch, serve & predict model eos8451 without issues. link to eos8451 ipynb file

Describe the steps to reproduce the behavior

To replicate:

• [x] Install Ersilia on Google Colab.
• [x] Connect Colab to the Drive, specify the path to the input file:
• [x] Convert your input file from csv to a smiles list to pass to Ersilia Models.
• [ ] Use Ersilia's Python Library to run predictions!

Expected behavior.

Model fetched successfully

Screenshots.

Operating environment

Windows 10. Brave browser version 1.44.105

Additional context

No response

[paulinebanye]

Update:

@GemmaTuron it's definitely an issue with the model. I just fetched eos6o0z without errors

[GemmaTuron]

Thanks for this @pauline-banye,

I have been able to get the model in my linux machine (see attached log: eos85a3.log )

And in this google colab as well: https://colab.research.google.com/drive/1LQtLcO3Vp-mCTxshP1n_5naO16KzJJ6h?usp=sharing Can you check if there are any pieces of code different ?

[Malikbadmus]

I'm also getting the same error on Model eos85a3.

This is the google colab link: https://colab.research.google.com/drive/142NLJo0Ho5HlG7_xKEkYYwdR1nSM-vUj?usp=sharing

[GemmaTuron]

I am really puzzled because running your Notebook @Malikbadmus I don't get any errors... is the first time running the same piece of code I get different outputs. I need to think a bit about what can be happening, open to suggestions as well!

[Malikbadmus]

Same here, Because the same block of code was used to successfully run two models.

Now I ran those two previously successful Models again and I'm getting the errors.

[paulinebanye]

Hmm @Malikbadmus I didn't experience the error affecting any of the other models. I fetched and predicted two models successfully after I encountered the error.

Thanks @GemmaTuron I'm comparing both notebooks now. I'll post an update once it's done

[Malikbadmus]

@GemmaTuron , I created a new notebook and I was able to successfully predict the model with the same line of code, so I figured maybe the open sessions were interrupting fetching new Models to run.

So I proceeded to Disconnect and delete the current runtime on the original notebook, and I was able to fetch the model successfully.

@pauline-banye, I wonder if you could try this solution and see if it works for you.

[paulinebanye]

@GemmaTuron I think this error is intermittent and possibly linked to the runtime as @Malikbadmus suggested. I compared your notebook to mine and the only difference was the verbose flag. I ran the predict again and it was successful.

Hi @Malikbadmus That's an interesting solution and I think you're probably right. When I reopened the notebook, I encountered a pop-up about reconnecting. I didn't delete the environment however but the end result is the same

[GemmaTuron]

That's very helpful thanks for the work @pauline-banye and @Malikbadmus ! See issue #378 I hope we will be able to understand where this is coming from, but meanwhile, would any of you add a comment on #359 so we can make sure users know what to do (at least temporarily) ?

Thanks !