11月12月1月2月3月4月5月6月7月8月9月10月
周一周三周五
一年内 256794 次贡献
更少更多
- frostbyte_neo 推送了仓库 github-featured/karpathy--llm.c 的 gordicaleksa-fix_dataloader2 分支比较 10 提交 »755458d0e1 fix tokenizer omg16635d41a2 attempt to fix PR9740a652cd Merge branch 'fix_dataloader2' of https://github.com/gordicaleksa/llm.c into gordicaleksa-fix_dataloader21787210306 Merge pull request #738 from ademeure/faster_compile68624cc6ac Refactor tokenize, delegate instead
- frostbyte_neo 推送了仓库 github-featured/karpathy--llm.c 的 fix/reshuffle_zero_grad 分支比较 10 提交 »7b92930021 move things around a bit, a few comments, and also update python script to correspond to C more8253823432 Merge branch 'fix_periodic_spikes' of https://github.com/gordicaleksa/llm.c into gordicaleksa-fix_periodic_spikes-2942fed59fe Merge pull request #662 from karpathy/overfit-batch-actually7fd7717816 fix the overfit single batch behavior to actually overfit batch, not microbatch, we do this more cleanly by simply resetting the dataloader every step74e5ba7260 Add micro step guard
- frostbyte_neo 推送了仓库 github-featured/dmlc--dgl 的 test_sparse_compile 分支比较 2 提交 »dd4335c342 [Sparse][Compile] Add GAT example2dee62b521 [Sparse] Support torchscript
- frostbyte_neo 推送了仓库 github-featured/karpathy--llm.c 的 fix/illegalaccess 分支比较 10 提交 »b9ad95e74c attempt to fix, but it's not working huh534b93a732 erase the old kernel to save LOC and do some renaming and adjust dev/cuda code to add the new largeC kernel as kernel 70554c7e414 Merge pull request #115 from ademeure/softmax_largeCfc3dce37ce Add __ldcs to a few reads for which the memory will never be read again in the forward pass321b042657 New softmax for non-self-attention softmax case with large Cs
