24/03/19 20:59:58.69 cXumgq0P.net
ビルドに2時間かかる
2:デフォルトの名無しさん
24/03/20 00:26:40.87 VbUXRLzi.net
LLVM IR出力すればPythonとかでもコンパイラ書けるのすごいと思う
3:デフォルトの名無しさん
24/03/25 19:29:53.92 3Ch+BeGW.net
Python + Larkで構文解析器して、llvmliteでLLVM IRを吐き出せば、誰でもコンパイラ作れる時代
4:デフォルトの名無しさん
24/04/23 10:43:47.75 2F7cqUxD.net
うちなんか、llvm -> lld -> compiler-rt -> clang+extra-tools までで5時間だぞ()
5:デフォルトの名無しさん
24/05/05 06:55:09.07 ieH+JJr4.net
おまえら、stand-alone build してないの?
6:デフォルトの名無しさん
24/06/11 12:24:03.99 6UMX2jU9.net
Let's reproduce GPT-2 (124M) - YouTube
URLﾘﾝｸ(www.youtube.com)
Andrej Karpathy
2024/06/10
We reproduce the GPT-2 (124M) from scratch. This video covers the whole process: First we build the GPT-2 network, then we optimize its training to be really fast, then we set up the training run following the GPT-2 and GPT-3 paper and their hyperparameters, then we hit run, and come back the next morning to see our results, and enjoy some amusing model generations. Keep in mind that in some places this video builds on the knowledge from earlier videos in the Zero to Hero Playlist (see my channel). You could also see this video as building my nanoGPT repo, which by the end is about 90% similar.
7:デフォルトの名無しさん
24/06/24 13:56:41.08 OzsKCRyN.net
加熱するLLM開発競争に冷や水、オープンモデルの組み合わせだけでGPT-4o越えの事実 ? WirelessWire News
URLﾘﾝｸ(wirelesswire.jp)
8:デフォルトの名無しさん
25/06/15 07:29:40.98 bNIMKunY.net
ヴァァージョンあげが頻繁すぎぃぃ