Development stopped?

[sgehrman]

Sorry I don't know much about what is going on. I use the backend to integrate into my app to do AI locally on my computer. Is this going away? What is this new olama REST API?

I don't see any commits for 2 months. Could someone explain what is going on?

[iwr-redmond]

#3542 contains commits less than a month old. If you need support for newer models while we all wait for some official word from Nomic AI about the future of this repo, you may wish to try the Nexa SDK. The instructions for starting a server that can receive API requests are here.

[sgehrman]

@iwr-redmond Tried Nexa SDK. Too buggy, development has halted. I thought AI was hot? Nothing works. Is there a real AI project that works. Nexa is cool if it wasn't buggy. What's similar to Nexa?

[iwr-redmond]

You could try Jan from Menlo Research.

[brankoradovanovic-mcom]

#3542 looks like a huge iteration. If that's what's going on at the moment, then it's no wonder there are no commits, and it's probably no cause for alarm either.

[PedzacyKapec]

@iwr-redmond Tried Nexa SDK. Too buggy, development has halted. I thought AI was hot? Nothing works. Is there a real AI project that works. Nexa is cool if it wasn't buggy. What's similar to Nexa?

You tried LM Studio ? They update their app sometimes twice a week :)

[TheAlex25]

Yeah, it seems that GPT4All is dead. If there are no updates by summer, it might be time to jump ship.

[vaultdweller-2287]

Unfortunately, the pull request describing switching to Ollama was.... rejected. Of course this doesn't 100% mean its over, they could just be iterating, but I wouldn't hold my breath.

From what I've seen there aren't many good alternative Python SDK's but I recommend trying llama-cpp-python. I haven't used it yet but it's on my list for an SDK to jump to.

[thingsiplay]

Looks like this project is dead? Last commit is 3 months ago. Sad, because this is the only user interface for me that can utilize my AMD GPU (the others just ignore it). And I liked the feature of creating my own custom Ai chatbot models with custom System Messages. Its easy to handle by just cloning an existing one.

This is my favorite LLM for Ai chat interface so far.

[PedzacyKapec]

Looks like this project is dead? Last commit is 3 months ago. Sad, because this is the only user interface for me that can utilize my AMD GPU (the others just ignore it). And I liked the feature of creating my own custom Ai chatbot models with custom System Messages. Its easy to handle by just cloning an existing one.

This is my favorite LLM for Ai chat interface so far.

Try LM Studio. THey also use Vulkan so in theory AMD should work there too.

[thingsiplay]

Looks like this project is dead? Last commit is 3 months ago. Sad, because this is the only user interface for me that can utilize my AMD GPU (the others just ignore it). And I liked the feature of creating my own custom Ai chatbot models with custom System Messages. Its easy to handle by just cloning an existing one.
This is my favorite LLM for Ai chat interface so far.

Try LM Studio. THey also use Vulkan so in theory AMD should work there too.

I actually did. The reason why I didn't want to use it is, because LM Studio is not open source. But my other alternatives (llamafile cli to web, and Alpaca) don't use my gpu either; only LM Studio and GPT4All did so far. I will use and play around with LM Studio and change to an open source alternative that works for me later; when I find one.

[vaultdweller-2287]

Update: The devs have returned? There’s a new commit from 16 hours ago updating “CircleCI”

[iwr-redmond]

This appears to be a fix for a critical security flaw. Hopefully @cebtenzzre can tell us whether GPT4all has entered maintenance mode, or whether core development will resume when resources permit.

[vaultdweller-2287]

I just E-mailed Nomic support about it. Will update!

[vaultdweller-2287]

Update: They responded with the following.

Thanks for getting in touch.

GPT4All is still an active project and continues to accept contributions.

More information about contributing can be found on the project's Github, and there's a Discord community for related conversation.

So they are continuing development... supposedly.

[iwr-redmond]

... supposedly.

Agreed.

You may wish to ask whether those acceptable contributions should be based on the main branch or the Ollama rewrite branch (#3542). If the correspondent has basic knowledge of the GPT4All development plan, I reckon they should be able to answer that question.