varlen

Here are 2 public repositories matching this topic...

conv attention natten varlen packed-tensor

Variable-length tensors with Triton kernels at 1.6× of FlashAttention-2 on H100. NumPy / PyTorch / MLX / JAX backends.

python machine-learning deep-learning gpu transformers pytorch triton primitive attention tensor mlx jax flash-attention ragged-tensor varlen

Add a description, image, and links to the varlen topic page so that developers can more easily learn about it.

To associate your repository with the varlen topic, visit your repo's landing page and select "manage topics."