Discussions

• 2 You must be logged in to vote

  📣

  Breaking & deprecation changes since v0.3

  eric-haibin-lin announced Apr 6, 2025 in Announcements

  

  

  

  

  3
• 1 You must be logged in to vote

  🙏

  How to load trained ckpt and eval with it on specific dataset?

  enhancement New feature or request good first issue Good for newcomers

  jankinf asked Feb 18, 2025 in Q&A · Answered

  

  

  

  

  6
• 0 You must be logged in to vote

  💡

  Understanding batch size configs in verl

  hiyouga started May 26, 2025 in Ideas / Show and tell

  

  

  0
• 1 You must be logged in to vote

  📝

  [RFC][sub roadmap][25Q2] Add Ascend NPU support for verl

  zheliuyu started Mar 31, 2025 in RFC

  

  

  2
• 1 You must be logged in to vote

  🙏

  Does the Verl framework support the base of the Moe architecture?

  orangelalala asked May 14, 2025 in Q&A · Answered

  

  

  

  

  2
• 2 You must be logged in to vote

  🙏

  Saving log to file without tee

  vadimkantorov asked May 12, 2025 in Q&A · Unanswered

  

  

  0
• 1 You must be logged in to vote

  📝

  [RFC] engine interface for training backends (FSDP, FSDP2, torchtitan, Megatron, PAI-Megatron, etc)

  RFC status: accepted

  eric-haibin-lin started May 2, 2025 in RFC

  

  

  4
• 1 You must be logged in to vote

  📝

  [RFC] verl recipe and extensions

  RFC status: accepted

  eric-haibin-lin started May 2, 2025 in RFC

  

  

  

  

  2
• 1 You must be logged in to vote

  📝

  [RFC] ActorEnvironment implementation to support tool calling

  good first issue Good for newcomers

  casper-hansen started Mar 6, 2025 in RFC · Closed

  

  

  

  

  20
• 1 You must be logged in to vote

  🙏

  Whether the verl framework supports qwen3

  orangelalala asked Apr 29, 2025 in Q&A · Answered

  

  

  1
• 1 You must be logged in to vote

  🙏

  GRPO roll out for one prompt

  hopekgh asked Mar 22, 2025 in Q&A · Answered

  

  

  1
• 1 You must be logged in to vote

  🙏

  Scheduling main_task on a worker that is not the head of ray cluster

  aashaka asked Apr 21, 2025 in Q&A · Answered

  

  

  1
• 1 You must be logged in to vote

  🙏

  I'm trying to understand the reason behind using Hydra as the internal configuration.

  zeocax asked Apr 13, 2025 in Q&A · Answered

  

  

  1
• 1 You must be logged in to vote

  📝

  Support synchronizing shard weights from training instance to inferencing instance

  hxdtest started Apr 16, 2025 in RFC

  

  

  0
• 1 You must be logged in to vote

  📝

  [RFC][Votable] Code Quality Tools Integration (Linter & Formatter & Import Organizer)

  RFC status: accepted

  tongyx361 started Apr 3, 2025 in RFC

  

  

  

  

  4
• 1 You must be logged in to vote

  🙏

  Issue with Multi-node training: failed to get register_center_actor

  EtashGuha asked Mar 5, 2025 in Q&A · Answered

  

  

  

  

  16
• 1 You must be logged in to vote

  🙏

  0.5B model OOM when Initializing actor_rollout

  MiaoLu3 asked Feb 6, 2025 in Q&A · Answered

  

  

  

  

  10
• 1 You must be logged in to vote

  🙏

  How to define custom compute_score function in v0.2

  souththzz asked Feb 18, 2025 in Q&A · Closed · Answered

  

  

  1
• 4 You must be logged in to vote

  📣

  v0.3.0.post0 release

  eric-haibin-lin announced Mar 30, 2025 in Announcements

  

  

  0
• 1 You must be logged in to vote

  📝

  [RFC] Add Ascend NPU support for verl

  Yikun started Feb 21, 2025 in RFC

  

  

  

  

  5
• 1 You must be logged in to vote

  🙏

  If the training process can resume from a checkpoint?

  takagi97 asked Feb 26, 2025 in Q&A · Closed · Unanswered

  

  

  1
• 1 You must be logged in to vote

  📝

  [RFC] Megatron-LM and MCore maintaining issues for veRL

  enhancement New feature or request megatron

  PeterSH6 started Nov 19, 2024 in RFC

  

  

  0