• 1
• 0
• 218

xcx_shine/omni-npu forked from omniai/omni-npu

A vLLM (0.12.0) out-of-tree platform plugin that enables running vLLM on NPU (Ascend/torch_npu).

• 1
• 0
• 544

xcx_shine/omni_infer forked from omniai/omni_infer

Omni_Infer is a suite of inference accelerators designed for the Ascend NPU platform, offering native support and an expanding feature set.

• 1
• 0
• 1380

xcx_shine/ModelZoo-PyTorch forked from Ascend/ModelZoo-PyTorch

暂停

• 1
• 0
• 0

xcx_shine/parallel decoding token-tree

用于并行解码，可把输入序列组织成树形式。