Closed mamabear25 closed 1 year ago
Hi @mamabear25 I need access to see the colab
And if you haven't, you can also add the -v flag in colab to get the error, when you use the !
it is the same as a CLI
Hi @mamabear25,
while you change access to colab, also check issue #378 as this might help!
@GemmaTuron you can access the colab now. I'll start trouble shooting right now Thanks
Same model isnโt successful on the CLI
Hi @mamabear25 I can't reproduce the CLI error, the screenshot you are showing is just the start of the fetching, so please save the log and let the fetch be completed. The outlined message is normal, it is simply checking that you dont have the model in the system
It doesnโt output anything after the last line on the screenshot, except i try terminating the process, then it'll continue fetching.
I'll have to control c
for it to continue
Is that normal?
Hi @mamabear25 I can't reproduce the CLI error, the screenshot you are showing is just the start of the fetching, so please save the log and let the fetch be completed. The outlined message is normal, it is simply checking that you dont have the model in the system
Hi @mamabear25
You should simply leave it running for a while and it will go on fetching. Please save the log file otherwise we wont know if there is an error
Okay, I'll do that
@GemmaTuron Here's the whole output
the traceback is similar to what I got on the google colab
@mamabear25 Have you tried deleting runtime and running all over again?
@Zainab-ik Yes, I have done that... same output
@GemmaTuron here's
I've been trouble shooting this model for days but nothing seems to work
It keeps outputing different errors on google colab, eg No module named dataclasses
,
Google colab
Same thing on the CLI
@GemmaTuron the other 3 models assigned me have been successful.
Can I leave model eos5qfo
and move forward?
Hi @mamabear25 !
Thanks for your efforts, yes, this is the same error as #387 for a different model. Can you change the issue name to: eos5qfo - Colab - no module named dataclasses
@GemmaTuron okay Thanks
Changing the miniconda installer from Miniconda3-py37... to Miniconda3-py39 seems to fix this issue. However not sure if this would break anything else / introduce regressions.
for reproducibility:
Indeed, using the Miniconda py3.9 solves the issue for this model, whereas using Miniconda py3.7 gives the following error at fetch time: ERROR: Could not find a version that satisfies the requirement contextvars; python_version < "3.7" (from bentoml) (from versions: 1.0, 2.0, 2.1, 2.2, 2.3, 2.4) ERROR: No matching distribution found for contextvars; python_version < "3.7"
15:21:09 | DEBUG | Error occurred while running: bash /tmp/ersilia-7na9ojia/script.sh > /tmp/ersilia-jylxbzaf/installs.log 2>&1 15:21:09 | ERROR | Ersilia exception class: ModelPackageInstallError
Detailed error: Error occured while installing package by running "bash /tmp/ersilia-7na9ojia/script.sh > /tmp/ersilia-jylxbzaf/installs.log 2>&1" command
Hi!
Quick update: With the fixes made by @carcablop on Google Colab (installing py37 again inside Colab) The model can be fetched without issues using the template modified here: https://github.com/GemmaTuron/ersilia/commit/b3f4d95f1f0be9ac190d2de9a65dce9bad7d8427
Describe the bug.
I'm getting a traceback error instead of the model being fetched succesfully.
Describe the steps to reproduce the behavior
Expected behavior.
๐ Model eos5qfo fetched successfully!
Screenshots.
Here's a link to my google collab
Operating environment
macOs Monterey Version 12.5.1
Additional context
No response