Day 0 support for Muse Spark Open on vLLM
·9 min read
Day-0 vLLM support for Muse Spark Open: 816B MoE, 1M context, multimodal inputs, DSpark speculative decoding, NVIDIA and AMD serving, initial TPU support, and LoRA RL.
#reinforcement-learning
8 posts
Finish RL Training Faster: Bitwise-Consistent, High-Performance, Agent-Driven
·35 min read
We make vLLM rollout and the training forward give bitwise identical logprobs on DeepSeek-V4-Flash and Nemotron-3.5-Lightning-30B-A3B (NVFP4), keep 91% of rollout throughput on DeepSeek-V4-Flash, and open-source the agent skill.
vime × RL-Kernel × AMD: Bitwise Train–Rollout Consistency on ROCm
·10 min read
vime and RL-Kernel align selected-token logprobs bit for bit across Megatron training and vLLM rollout on AMD Instinct MI300X, with zero mismatches across 200 GRPO steps.
Large-Scale Sharded Weight Transfer with Ray Direct Transport (RDT) in vLLM
·18 min read
We implement a native sharded weight transfer engine in vLLM utilizing Ray Direct Transport (RDT), achieving weight transfer for the Kimi K2 model in BF16 on 48 8xH100 nodes in 7.53s
IsoExec: Unified Execution to Eliminate Trainer-Inference Mismatch in SkyRL
·11 min read
IsoExec unifies numerical execution across SkyRL's vLLM and Megatron runtimes, reducing the average rollout-versus-training logprob difference below 1e-6 on Qwen3.5-35B-A3B with 25% overhead.
vime + ROCm: End-to-End RL Post-Training on AMD Instinct™ GPUs
·10 min read
Announcing ROCm support for vime, now running end-to-end on AMD Instinct MI355X GPUs with prebuilt container.
Announcing vime: A Simple, Stable, and Efficient RL Framework for LLMs
·6 min read
vime connects slime's training stack with vLLM rollouts to provide a simple, stable, and efficient RL post-training pipeline.
Native RL APIs in vLLM
·12 min read
How vLLM native RL APIs standardize weight syncing and asynchronous RL serving with NCCL and CUDA IPC transfer backends, pause mode, and fixes for fragile DPEP and disaggregated rollout deployments.