dhcode-cpp / easy-dualpipeLinks
Pipeline-Parallel Lecture: Simplest Dualpipe Implementation.
☆25Updated last month
Alternatives and similar repositories for easy-dualpipe
Users that are interested in easy-dualpipe are comparing it to the libraries listed below
Sorting:
- simplest online-softmax notebook for explain Flash Attention☆13Updated 7 months ago
- ☆140Updated last month
- Implementation of FlashAttention in PyTorch☆159Updated 6 months ago
- ☆145Updated 5 months ago
- Efficient, Flexible, and Highly Fault-Tolerant Model Service Management Based on SGLang☆55Updated 9 months ago
- qwen-nsa☆70Updated 3 months ago
- Triton Documentation in Chinese Simplified / Triton 中文文档☆78Updated 3 months ago
- A flexible and efficient training framework for large-scale alignment tasks☆400Updated this week
- An easy-to-use package for implementing SmoothQuant for LLMs☆103Updated 4 months ago
- Tiny-DeepSpeed, a minimalistic re-implementation of the DeepSpeed library☆41Updated last week
- from MHA, MQA, GQA to MLA by 苏剑林, with code☆25Updated 5 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆134Updated 3 weeks ago
- [ICLR 2025] PEARL: Parallel Speculative Decoding with Adaptive Draft Length☆102Updated 3 months ago
- ☆52Updated 2 years ago
- A MoE impl for PyTorch, [ATC'23] SmartMoE☆66Updated 2 years ago
- an implementation of transformer, bert, gpt, and diffusion models for learning purposes☆155Updated 9 months ago
- ☆52Updated this week
- llm theoretical performance analysis tools and support params, flops, memory and latency analysis.☆101Updated 3 weeks ago
- Inference code for LLaMA models☆122Updated last year
- DeepSeek Native Sparse Attention pytorch implementation☆83Updated 5 months ago
- Transformer related optimization, including BERT, GPT☆17Updated 2 years ago
- 使用单个24G显卡,从0开始训练LLM☆56Updated last month
- [USENIX ATC '24] Accelerating the Training of Large Language Models using Efficient Activation Rematerialization and Optimal Hybrid Paral…☆61Updated last year
- ☆112Updated 8 months ago
- ☆79Updated last year
- A LLaMA1/LLaMA12 Megatron implement.☆28Updated last year
- Implement some method of LLM KV Cache Sparsity☆35Updated last year
- (best/better) practices of megatron on veRL and tuning guide☆41Updated last week
- 使用 CUDA C++ 实现的 llama 模型推理框架☆58Updated 9 months ago
- Train llm (bloom, llama, baichuan2-7b, chatglm3-6b) with deepspeed pipeline mode. Faster than zero/zero++/fsdp.☆97Updated last year