liangyuwang / Tiny-DeepSpeedLinks
Tiny-DeepSpeed, a minimalistic re-implementation of the DeepSpeed library
☆49Updated 5 months ago
Alternatives and similar repositories for Tiny-DeepSpeed
Users that are interested in Tiny-DeepSpeed are comparing it to the libraries listed below
Sorting:
- Efficient Mixture of Experts for LLM Paper List☆166Updated 4 months ago
- Implementation for FP8/INT8 Rollout for RL training without performence drop.☆289Updated 3 months ago
- ☆209Updated 3 months ago
- qwen-nsa☆87Updated 3 months ago
- [ICLR 2025] PEARL: Parallel Speculative Decoding with Adaptive Draft Length☆147Updated last month
- A lightweight reinforcement learning framework that integrates seamlessly into your codebase, empowering developers to focus on algorithm…☆101Updated 5 months ago
- Tiny-Megatron, a minimalistic re-implementation of the Megatron library☆21Updated 5 months ago
- DeepSeek Native Sparse Attention pytorch implementation☆115Updated last month
- 青稞Talk☆190Updated 2 weeks ago
- Rethinking RL Scaling for Vision Language Models: A Transparent, From-Scratch Framework and Comprehensive Evaluation Scheme☆147Updated 10 months ago
- Pipeline-Parallel Lecture: Simplest Dualpipe Implementation.☆31Updated 4 months ago
- ☆152Updated 7 months ago
- Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs☆204Updated 2 months ago
- ☆68Updated last year
- [ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Training☆258Updated 6 months ago
- MiroRL is an MCP-first reinforcement learning framework for deep research agent.☆229Updated 5 months ago
- ☆129Updated 8 months ago
- Bridge Megatron-Core to Hugging Face/Reinforcement Learning☆191Updated last week
- ☆87Updated 5 months ago
- ☆230Updated last month
- A set of examples based on verl for end-to-end RL training recipes.☆152Updated last week
- SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs☆192Updated 4 months ago
- Official Implementation of "Learning Harmonized Representations for Speculative Sampling" (HASS)☆54Updated 10 months ago
- A highly capable 2.4B lightweight LLM using only 1T pre-training data with all details.☆223Updated 6 months ago
- Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning☆193Updated 10 months ago
- A visuailzation tool to make deep understaning and easier debugging for RLHF training.☆284Updated 11 months ago
- Skywork-MoE: A Deep Dive into Training Techniques for Mixture-of-Experts Language Models☆139Updated last year
- TransMLA: Multi-Head Latent Attention Is All You Need (NeurIPS 2025 Spotlight)☆429Updated 4 months ago
- L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning☆261Updated 8 months ago
- ☆87Updated last week