Discussions

• 1 You must be logged in to vote

  🙏

  How Ring Attention Works to Qwen3.5? Is It Allowed?

  fahadh4ilyas asked May 5, 2026 in Q&A · Unanswered

  

  

  0
• 1 You must be logged in to vote

  🙏

  请问一下，actor的初始化里dschf为什么创建了却没有使用呢？

  rangehow asked Feb 25, 2025 in Q&A · Unanswered

  

  

  1
• 1 You must be logged in to vote

  🙏

  Question about weight update mechanism.

  The-Hierophant asked Jan 26, 2025 in Q&A · Unanswered

  

  

  1
• 1 You must be logged in to vote

  💡

  Add Gradient Accumulation logic in training code to enhance training consistency

  ZetangForward started Oct 26, 2024 in Ideas · Closed

  

  

  2
• 1 You must be logged in to vote

  💬

  process groups for actor and vllm engines

  babu111 started Jul 12, 2024 in General · Closed

  

  

  0
• 2 You must be logged in to vote

  🙏

  What hardware is required?

  EUTW06 asked Mar 16, 2024 in Q&A · Unanswered

  

  

  0
• 1 You must be logged in to vote

  🙏

  Using tensorboard instead of wandb

  pratikkumar018 asked Mar 5, 2024 in Q&A · Answered

  

  

  7
• 1 You must be logged in to vote

  💬

  Difference between micro_train_batch_size & micro_rollout_batch_size

  pratikkumar018 started Feb 16, 2024 in General

  

  

  

  

  4
• 1 You must be logged in to vote

  💬

  Error in PPO when using resized model

  pratikkumar018 started Jan 30, 2024 in General

  

  

  10
• 1 You must be logged in to vote

  📣

  We have remove the pad_token for llama2

  hijkzzz announced Dec 17, 2023 in Announcements · Closed

  

  

  0
• 1 You must be logged in to vote

  📣

  Fix pad_token overflow in word_embedding

  hijkzzz announced Aug 27, 2023 in Announcements · Closed

  

  

  0
• 1 You must be logged in to vote

  📣

  Pre-trained SFT/RM checkpoint

  hijkzzz announced Aug 16, 2023 in Announcements · Closed

  

  

  0