Reflection-Llama-3.1-70B on the ProLLM Coding Assistant benchmark

Transformer and Human Brain - Andrej Karpathy ----- Video Credit - Original video from "No Priors: AI, Machine Learning, Tech, & Startups" YouTube Channel (Link in comment)

Linkedin's great Liger Kernel repo just released the updated version. - New Integrations: SFTtrainer, Axolotl, LLaMa-Factory - New Models Support: Phi3 & Qwen2 - AutoModel API: Meet AutoLigerKernelForCausalLM - Enhanced FusedLinearCrossEntropy: support bias term

Andrej Karpathy on the importance of extremely smaller-sized distilled models (even 1Bn param model should be good enough) Video Credit - Original video from "No Priors: AI, Machine Learning, Tech, & Startups" YouTube Channel (Link in comment)

This 1.58 bit LLMs with llamafile is so interesting. - Weights: -1, 0, +1 (like Soviet Setun) - 8 size options: 78MB to 1.3GB - Strong benchmark performance on 4 CPUs -------- llamafile is a new format introduced by Mozilla Ocho on Nov 20th 2023. It uses Cosmopolitan Libc to

Nice Github project - Train FLUX LoRA locally with LOW VRAM (12GB/16GB/20GB) and with UI

😅🤣

👀

"OpenAI clarifies: No, "GPT Next" isn't a new model."

The 100k H100 cluster of xAI has begun training