Past roadmap dicusssions for reference: #710 #22

The most important thing for verl Q3 is to make it a modular foundational library for the community to extend, as a starting point but not the destination.

composable model engines

Finish up #1560 such that parallelism strategy is not implemented at the engine level, without exposing details to the worker(role) level. The fsdp/megatron engines are expected to be created and run in a standalone fashion, and be reused across different roles.

• fsdp actor, critic, ref (focus on fsdp2)
• megatron actor, critic, ref
• torchtitan integration (call for contribution)
• switch all recipe/examples from fsdp1 to fsdp2 by default (and remove ill-maintained ones)

Work in progress interface for comments #1977

rollout workers

• optimize server mode rollout performance
• modular rollout workers: VllmRolloutWorker and SGLangRolloutWorker, exposing the same APIs
• support model with random init weight
• weight resharding: optimize tp x dp dispatch, and support receiving weight from separate resource groups
• Agent RL infrastructure [agentic RL] multi-turn rollout and agent loop development tracking #2618

Additional ongoing efforts:

• Multi-turn rollout & agentic RL Status & Roadmap zhaochenyang20/Awesome-ML-SYS-Tutorial#131
• [roadmap] Rollout Module Development Progress & Roadmap #1882

async & disaggregated architecture

• one-step off async pipeline (WIP: [trainer, fsdp, vllm, recipe] feat: one step off async training recipe #2231), further performance optimization & profiling needed
• streaming/partial rollout (WIP: [rollout] feat: support reorder rollout for tackling long-tail generation problem #2200)
• performance tuning, and reference throughput benchmark across [model type, model size, seqlen, hardware, num accelerators, worker role] to achieve better disaggregated resource allocation
• fully-async pipeline

multi-turn, data, config infra

• better message infra for multi-turn messages, dense reward @SwordFaith
• better dataset schema for train & rollout. We need documentation too. TRL's documentation is good https://huggingface.co/docs/trl/en/dataset_formats @SwordFaith
• use tensordict and nested-tensor to remove padding and replace DataProto
• replace omegaConfig with read-only dataclass for verl internal config passing [cfg] refactor: make actor config more modular #2379 https://github.com/volcengine/verl/pull/2147/files and make unit test easier
• P1: distributed data pool from https://arxiv.org/pdf/2507.01663v1 [RFC] Add persistable replay buffer for large-scale rollout data storage #2539

streamline new model workflow

• document the workflow to add a new hf model to verl. Currently with latest vllm there's no need to add weight loader mentioned in https://verl.readthedocs.io/en/latest/advance/fsdp_extension.html
• better abstraction and registration system for multi-modal models. Currently different multi-modals have inconsistent config attr (e.g. rope), freeze/unfreeze setup, input/output processing... (ideally this should be done at huggingface transformers level but it's not sufficient right now cc @NielsRogge) (RFC needed)
• verl needs a documentation page about the latest status of model support and per model related features (lora, sequence parallelism, megatron, etc)

high quality recipes and end2end optimizations

• retool recipe (code is ready, going through reviews)
• SOTA multimodal vlm RL recipe (call for contribution)
• enhance DAPO recipe with larger models, and provide scripts with high training throughput (many perf knobs are not turned on in the current script)
• we welcome more recipes from the community, please open an RFC if you're interested in contributing before opening any PR for recipes [RFC] Verl recipe for image generation model #2136

Additional existing ongoing features:

• [mcore] verl+megatron development tracking #1033
• Features that npu will focus on supporting in Q3 #2171

Many roadmap tasks in this doc are initiated by & credit to @vermouth1992 @SwordFaith