github.com
GitHub - JustVugg/nanoeuler: GPT-2-style LLM built from scratch in C/CUDA with hand-written backprop, BPE tokenizer, FlashAttention, pretraining, and SFT. · GitHub
GPT-2-style LLM built from scratch in C/CUDA with hand-written backprop, BPE tokenizer, FlashAttention, pretraining, and SFT. - JustVugg/nanoeuler
If the page stays blank, open it in a new tab. Your Weird rating still works from the top bar.
Open source