A reading list from Attention Is All You Need to FlashAttention drew about 731 likes.
@cneuralnetwork shared nine papers to read, in order, for learning how large language models work.
The list: Attention Is All You Need, BERT, GPT-2, GPT-3, Chinchilla, InstructGPT, LoRA, LLaMA and FlashAttention.
Why it mattersIt covers the main ideas behind modern models, from the transformer to scaling, instruction tuning, cheap fine-tuning and faster attention.
One 30B MoE checkpoint handles transcription in 40 languages, speaker attribution and translation.
@NVIDIAAI2.5k
A Qwen ambassador used the model's state as memory so it can run C and MicroPython.
@QwenDevs2.4k
Ayush Agarwal says it wasn't a hack or a GitHub bug. About 1,025 likes.
@ayushagarwal1.5k
Non-profit Murai Labs released base, chat and GGUF files under Apache-2.0.
@Mechramc1.2k
It ships a CLI, linter, UI, LSP and agent integrations. About 797 likes.
@antfu71.1k
The model runs with 2ms latency on an RTX 5090, according to the post.
@kimmonismus888One email each morning with the AI, crypto and tech posts everyone is talking about, explained in plain English.
