Pinned Loading

1. Megatron-Energon Megatron-Energon Public

   Forked from NVIDIA/Megatron-Energon

   Megatron's multi-modal data loader

   Python

2. flash-attention flash-attention Public

   Forked from Dao-AILab/flash-attention

   Fast and memory-efficient exact attention

   Python

3. Megatron-LM Megatron-LM Public

   Forked from NVIDIA/Megatron-LM

   Ongoing research training transformer models at scale

   Python

4. TransformerEngine TransformerEngine Public

   Forked from NVIDIA/TransformerEngine

   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

5. MMseqs2 MMseqs2 Public

   Forked from soedinglab/MMseqs2

   MMseqs2: ultra fast and sensitive search and clustering suite

   C

6. diamond diamond Public

   Forked from bbuchfink/diamond

   Accelerated BLAST compatible local sequence aligner.

   C++