• 1
• 0
• 0

王思源/MixFrame

支持continuous batching与disaggregate两种batching模式的推理框架

• 1
• 0
• 0

王思源/tmp_swiftransformer

加了profiling功能的swifttransformer

• 1
• 0
• 0

王思源/mydistserve

基于distserve的更改

• 1
• 0
• 0

王思源/SwiftTransformer

更新了flash attention的原版swifttransformer

• 1
• 0
• 0

王思源/vllm_flash_attention

• 1
• 0
• 0

王思源/composable_kernel

copy of composable_kernel

• 1
• 0
• 0

王思源/timing-sglang

record sglang

• 1
• 0
• 0

王思源/tmpvllm

tmp vllm for server

• 1
• 0
• 0

王思源/my-pictures

我的图片库，用于加载网络图片

• 1
• 0
• 37

王思源/FBDQA-2023S forked from saturnlab/FBDQA-2023S

Financial Big Data and Quantitative Analysis

• 1
• 0
• 0

王思源/SwiftTransformerDeveloped

adapted from SwiftTransformer,added the load of Llama3 qwen3 series models,profiling function