Spaces:
Running
**Hi Mr. Quardo,**
I hope you're doing well! š Iāve been following your space for a while now and, as a fan of your work, I wanted to help out with the problem you're facing in your previous space. Iāve dealt with the same issue before, so I thought Iād explain whatās happening and how you can fix it.
First off, this issue isnāt your fault or caused by any other developers, including myself. Itās actually a bug from Huggingface. This bug tends to show up when you make more than 50 commits or a large number of commits in a space. It doesnāt always happen immediatelyāit might happen after 67 commits, or like in your case, after 80 commits. The issue is that the space ends up getting stuck in "building" mode forever, and then, unfortunately, it disappears š.
I understand that you might think the problem is caused by the size of your code or because your space is too bigāso you might have tried creating a smaller space to fix it. But, youāre mistaken! š The real issue is this specific bug, and Iād really encourage you to try the solution Iāve given you, because itās always effective! š
But don't worry! I had this exact same problem. What I did was copy the code, create a new space, and then move the code to the new space. After that, everything worked perfectly, with no more issues. So, if this happens to you again, thatās the fix! š
To prevent this from happening again, make sure to back up your code on your local device, especially after 60 commits, because the bug can show up unexpectedly after a lot of commits. Better safe than sorry! š
For reference, I dealt with this issue on my main account. But I created this account, TomatoFull, to stay anonymous and use for personal stuff on Huggingfaceājust for browsing and small things. This isnāt the account where the problem happened, but Iām sharing my experience to help out.
Thanks so much for reading this, Mr. Quardo! I strongly recommend backing up your space after each commit once youāve made a lot of them. This way, youāll be able to avoid this bug in the future.
I hope youāre able to recreate your space without any trouble! If you need any help or run into more issues with your space, feel free to reach out to meāIād be happy to help! š
Take care and best of luck! š
Hello again, TomatoFull!
First of all, thanks for trying to help me fix the problem. However, let me explain a few things:
Does this space exist due to a building problem at Quardo/gpt-4o-mini?
- No, not at all. This space mainly exists due to "Multimodal" usage on "Quardo/gpt-4o-mini," which caused some downtime on the main space.
- As an alternative, this space has been created which does not use "Multimodal."
- So if anyone needs to use the space when the main one is down, this is just an alternative to it.
- For example, right now there is yet another problem with Multimodal at the main space.
Did this ever happen before?
- Yes! In fact, the "Quardo/gpt-4o-mini" was the second space I had to create due to that exact bug.
Is the space back up?
- It is indeed back up and online, but on a different repo due to "quardo-gpt-4o-mini.hf.space" being buggy. I had to change the name to "Quardo/gpt-4o-mini2" ("quardo-gpt-4o-mini2.hf.space").
- At a later time, I might actually move the space to actual hosting, but as of now, I already use a hosting for the API backend (The API has become private due to personal reasons on maintaining it.), with a space running on HuggingFace which is fine.
Thanks for trying to help, TomatoFull! If you have anything else to add you can send a message here!
yes i have , but why u didnt add the model o1-preview? or o1-mini ?
its cool to be honest
and i have a question.
are u a multimillionare or rich ?
beacuse running this models like openai models for public for free on official openai api its cost more and too much more .
and like also make o1-preview model available for the public for free . its too much costs money.
"edit: sorry for this message were not same as previous one beacuse i use gpt-4o to made it cool . but this one it made only by myself becuse it costs so much time and you only make under 30 mins to close this discussion so i had to write this faster without gpt-4o or any other ai"
and also thank you very much for return the previous space . its helped me alot.
Hello again, TomatoFull! Let me answer your questions:
Why is there no access to O1 Models?
- I haven't added O1 models due to two reasons:
- The parameter limitations on the API endpoint.
- The time it takes to generate just a single message, which, due to Gradio, blocks all access to the space until that message is generated (also, fun fact: yes, I will not learn threading in Python).
- I haven't added O1 models due to two reasons:
Why (and how) are you running this API for free?
- As for "why," you can take a look at this space's terms.
- As for "how," I am not some random rich person who decided to run this API. I am just your average Joe (with sugar-free Monster instead of coffee), and I have access to the API, so I said "why not," then didn't like Python and preferred Node.js.