
Guides long-context MoE training in Megatron Bridge: CP sizing, recompute and dispatcher
- Open source
- 4K GitHub stars
- AI models and LLM tooling


Guides long-context MoE training in Megatron Bridge: CP sizing, recompute and dispatcher

Runs Megatron-LM and Megatron Bridge training and checks their loss curves match

Guides choosing and combining TP, DP, PP, CP and EP degrees for Megatron Bridge training

Guides an agent through enabling Megatron FSDP sharded training in Megatron-Bridge

Guides an agent through enabling hierarchical context parallelism in Megatron-Bridge

Picks Megatron Bridge training recipes and resizes parallelism for your GPU allocation

Hardware-specific MoE training playbooks and candidate parallelism configs for Megatron

Guides an agent through reducing peak GPU memory in Megatron Bridge training runs

Guidance for choosing FSDP or 3D parallelism when training MoE vision-language models

Configures CPU offloading in Megatron Bridge to cut GPU memory during training

Guides tuning of MoE expert-parallel communication overlap in Megatron Bridge

Guides enabling and verifying TP, DP and PP communication overlap in Megatron-Bridge
The closest alternatives to Nemo mbridge perf sequence packing are Nemo mbridge perf moe long context, Nemo mbridge mlm bridge training, Nemo mbridge perf parallelism strategies, Nemo mbridge perf megatron fsdp and Nemo mbridge perf hierarchical context parallel.
Yes. Nemo mbridge perf moe long context, Nemo mbridge mlm bridge training, Nemo mbridge perf parallelism strategies have a free or open-source option.