Search results

Question about TTRL training schedule: reward convergence & test data usage

Status: Open.

#48 In PRIME-RL/TTRL;

· tls0523 opened on Dec 3, 2025

Environment conflict assistance

Status: Open.

#47 In PRIME-RL/TTRL;

· user683 opened on Nov 16, 2025

Unexpected Performance for Qwen2.5-7B-MAth

Status: Open.

#46 In PRIME-RL/TTRL;

· BodongDu opened on Nov 16, 2025

How can I reproduce the results for PPO or PRIME?

Status: Open.

#45 In PRIME-RL/TTRL;

· ZhaoningYu1996 opened on Nov 14, 2025

Request for evaluate code

Status: Open.

#44 In PRIME-RL/TTRL;

· Marcowky opened on Nov 6, 2025

AssertionError: ppo_mini_batch_size 0 should be larger than 0 after normalization when running examples/ttrl/Qwen2.5-Math/math.sh

Status: Open.

#43 In PRIME-RL/TTRL;

· childofcuriosity opened on Oct 25, 2025

LoRA support?

Status: Open.

#42 In PRIME-RL/TTRL;

· soumyasj opened on Oct 15, 2025

Question for selection of experiments results

Status: Open.

#40 In PRIME-RL/TTRL;

· Trae1ounG opened on Sep 29, 2025

The new code dose not include a reward function for gpqa？

Status: Open.

#39 In PRIME-RL/TTRL;

· mootutu opened on Sep 28, 2025

Requests for gpqa.sh

Status: Open.

#38 In PRIME-RL/TTRL;

· ValoraY opened on Sep 9, 2025

Where in the code is it shown that the reward model is also being trained?

Status: Open.

#34 In PRIME-RL/TTRL;

· nikiniki0127 opened on Jul 31, 2025

The process freezes after creating a Ray cluster (while trying to follow the Getting Started guide).

Status: Open.

#32 In PRIME-RL/TTRL;

· AIgamatrix opened on Jul 15, 2025