So #Steeve got a major upgrade recently. He moved from a #gptneo (2.4B) model to a #llama2 (7B) model. Trained on 300k messages from our private chat history, Steeve is way more capable of following the conversation now. He used to have some "favorite phrases" he would say a lot, and I'm seeing less of that. His vision and reading models also got upgraded, so he gets more detail about the links and memes we share. Long live Steeve! 

#ai #chatbot #llm #llama #gpt #transformers

The future of generative AI is niche, not generalized

By better bridging the gap between generative AI and more specific and niche datasets, over time people should build a subtly different relationship with the technology. It will lose its mystique as something that ostensibly knows everything, and it will instead become embedded in our context.

#artificialintelligence #AI #GenerativeAI #LLM #GPTJ #GPTNeo #technology #tech #innovation

https://www.technologyreview.com/2023/04/27/1072102/the-future-of-generative-ai-is-niche-not-generalized/

 #Steeve having a normal one tonight.

#ai #gpt #gptneo

Hey #ai geniuses, I've been fine tuning #gpt2 and #gptneo models for a while with, but my graphics card being what it is (and my training corpuses being *huge*) I would like to train a nice midsize model. Something bigger than their 125M, but something smaller than their 1.3B model. I've had zero success getting anything working when applying my training scripts to the #bloom 560M model. Loss converges to zero almost instantly. Got any experience to share?

Please #boost for visibility plz

More fun from #Steeve the #AI #chatbot. For context, he was recently upgraded with the instruct dataset from the alpaca project, specially modified to include instructions on how to respond like us. This is his reaction to the command "!instruct how to sex properly" 

#gpt #gpt2 #gptneo

My current main objection towards #OpenAI is that it is not #FreeSoftware; and from that I am thinking of checking out #GPTNeo, which is.

Edit: I have seen an objection to this view; data about un-sentient beings are different from data about sentient beings. See more: https://vimeo.com/281704944

I finally might have found a #MachineLearning task that the Apple M1 machine does better than the PC with the RTX 3090 …

I wanted to fine-tune the GPT Neo 2.7 Billion model. Most docs/articles say not to. But I wanted to 😛 While I only ran 4 steps, the M1 did it in about 20 minutes or so (I think) last night.

I’ve had the PC running the same task today and even after an hour or so, it seems not have even started … but need to wait and see if something happens.

#DeepLearning #GPTNeo

⭕ Stable Diffusion - IA OpenSource de Generación de imágenes con tu tarjeta gráfica.

➡️ https://t.me/aitorroma/1014

#lifehacks #ia #opensource #gptneo #graphics