Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
Back to feed
Nitral-AI 
posted an update 2 days ago
Post
2819
That moment when you spend 5 days up babysitting trains, only for colab pro + to randomly disconnect the environment at every chance with 0 error indication of any kind (it just disconnects without an error). Nuke the session from the interface, but continue to eat my colab credits while it reports to wandb. 0 way of saving the models when this happens since it nukes the code preset up to auto-execute. And since the sessions 'exist' but also at the same time doesn't exist i cant close it. And have to wait till they auto timeout after 24hrs. Guess, i won't be using colab for 'quick' test trains anymore. Thanks google for scheming the very little model training budget i had for the month.
Nitral-AI
•
edited 1 day ago

After an extended interaction with support regarding the issue, the experience has been nothing short of frustrating:

They managed to misidentify my subscription tier twice within the same email thread (quite the achievement).
Repeatedly suggested I purchase more credits, even though my issue had/has absolutely nothing to do with credits (a fantastic way to build trust and totally not come off as scammy).
It turns out Colab offers no meaningful customer support beyond payment-related issues, rendering them entirely useless for technical insights via email [couldnt find a phone number that worked]. (They seem primarily concerned with ensuring they can take your money without issue.)
Their recommended external support options—GitHub, Stack Overflow, or the internal Colab notebook help—are dead ends. The GitHub issues page and related Stack Overflow threads are overrun with unresolved and unexplained disconnection problems. (For the record, I’ve submitted plenty of bug reports myself in note-book—nothing seems to change, no reply or follow up of any kind, behavior remains persistent. The github issues page doesn't really improve my sense of confidence in them fixing this either.)

·
Lewdiculous

They managed to misidentify my subscription tier twice within the same email thread (quite the achievement).

Repeatedly suggested I purchase more credits, even though my issue had/has absolutely nothing to do with credits (a fantastic way to build trust and totally not come off as scammy).

Holy, this is unacceptable support quality for a paid service, not to mention the issue itself directly harming you by burning credits. Shameful.

In this post