Congrats! I went to your demo and asked for words that end in agi. This is what I got:<p>--<p>agi, agi, agi, agi, agi, agi, agi<p>These are some of the words that end in agi. You can also use the word agi in a sentence. For example, "I am going to the grocery store to get some agi."<p>These are some of words that end in agi.<p>These are some words that end in agi.<p>maximize, maximize, maximize, maximize, maximize, maximize, maximize, maximize<p>These are some words that ends in agi<p>--<p>So I think this needs more work to get to "as good as ChatGPT". But having said that, congrats on the landing
Hi HN!<p>I’m super excited to announce Lamini, the LLM engine that gives every developer the superpowers that took the world from GPT-3 to ChatGPT!<p>I’ve seen a lot of developers get stuck after prompt-tuning for a couple days or after fine-tuning an LLM and it just gets worse—there’s no good way to debug it. I have a PhD in AI from Stanford, and don’t think anyone should need one to build an LLM as good as ChatGPT. A world full of LLMs as different & diverse as people would be even more creative, productive, and inspiring.<p>That’s why I’m building Lamini, the LLM engine for developers to rapidly customize models from amazing foundation models from a ton of institutions: OpenAI, EleutherAI, Cerebras, Databricks, HuggingFace, Meta, and more.<p>Here’s our blog announcing us and a few special open-source features! <a href="https://lamini.ai/blog/introducing-lamini" rel="nofollow">https://lamini.ai/blog/introducing-lamini</a><p>Here’s what Lamini does for you:
Your LLM outperforms general-purpose models on your specific use case
You own the model, weights and all, not us (if foundation model allows it, of course!)
Your data helps the LLM, and build you an AI moat
Any developer can do it today in just a few lines of code
Commercial-use-friendly with a CC-BY license<p>We’re also releasing several tools on Github:
Today, you can try out our hosted data generator for training your own LLMs, weights and all, without spinning up any GPUs, in just a few lines of code from the Lamini library. <a href="https://github.com/lamini-ai/lamini/">https://github.com/lamini-ai/lamini/</a><p>You can play with an open-source LLM, trained on generated data using Lamini. <a href="https://huggingface.co/spaces/lamini/instruct-playground" rel="nofollow">https://huggingface.co/spaces/lamini/instruct-playground</a><p>Sign up for early access to the training module that took the generated data and trained it into this LLM, including enterprise features like virtual private cloud (VPC) deployments. <a href="https://lamini.ai/contact" rel="nofollow">https://lamini.ai/contact</a>
This headline is totally editorializing. Stick with the source one. “Introducing Lamini, the LLM Engine for Rapidly Customizing Models”<p>So much click bait in the LLM space.
The actual post doesn't say "as Good as ChatGPT", why does the HN title?<p>I don't really care to click on something I know is obviously lying to me.
I've been playing a bit with stacking transformer adapters to add knowledge to models and so far it has met my needs. It doesn't have the same illusion of intelligence, but so far it's just as good as a multitasking intern, so I am still having fun with it. I wonder if this is basically doing the same thing.
GPT at this point is more than an LLM, it is a baseline layer of logic using the underlying transformer technology. This will be challenging to replicate without the same size of data sets
Noting that the Github repo includes a data pipeline for instruction fine tunining.<p>What's the difference between this and other data pipelines like Alpaca?