Pinned Loading

1. bert_from_scratch bert_from_scratch Public

   bert-inspired model

   Python 1

2. Triton-guide Triton-guide Public

   A hands-on guide to writing fast custom GPU kernels in Triton :-- from your first vector addition to fused attention mechanisms.

   3

3. NVIDIA/Megatron-LM NVIDIA/Megatron-LM Public

   Ongoing research training transformer models at scale

   Python 18.2k 4.6k

4. sparseFlow sparseFlow Public

   High-performance sparse attention kernels and distributed routing primitives for modern LLM systems.

   Python 1

5. deepseek-v3 deepseek-v3 Public

   Deepseek-v3 inspired: scaled-down implementation with single GPU training

   Python

6. NVIDIA/TransformerEngine NVIDIA/TransformerEngine Public

   A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on Hopper, Ada and Blackwell GPUs, to provide better performance…

   Python 3.6k 867