Search results

Constant scoring in step_norm_reward function in core_gigpo.py

Status: Open.

#217 In langfengQ/verl-agent;

· Shirley-Kokane opened on Dec 27, 2025

Training Process can not start, may be an issue with multiple processes.

Status: Open.

#216 In langfengQ/verl-agent;

· y1501028421 opened on Dec 27, 2025

SearchR1 server issue

Status: Open.

#215 In langfengQ/verl-agent;

· GongYiPu opened on Dec 24, 2025

How to verify the trained VLM model and visualize the environment interaction?

Status: Open.

#213 In langfengQ/verl-agent;

· Jack1447 opened on Dec 22, 2025

How to iterate over the full training dataset in small batches instead of launching thousands of parallel processes?

Status: Open.

#212 In langfengQ/verl-agent;

· Celoond opened on Dec 18, 2025

web search model weights

Status: Open.

#211 In langfengQ/verl-agent;

· jos-fisher-jejgkej opened on Dec 16, 2025

Question: Performance difference between the [bamboogle/test_score] and [val/bamboogle_success_rate]

Status: Open.

#210 In langfengQ/verl-agent;

· hanzhou032 opened on Dec 16, 2025

Logic error in the function print_size in protocol.py

Status: Open.

#205 In langfengQ/verl-agent;

· whiteHuskie opened on Nov 30, 2025

Training ALFWorld with small number of CPUs

Status: Open.

#204 In langfengQ/verl-agent;

· hosytuyen opened on Nov 29, 2025

Training Questions related to ALFWorld Reproduction

Status: Open.

#203 In langfengQ/verl-agent;

· yanx1326 opened on Nov 25, 2025

Image features and image tokens do not match

Status: Open.

#202 In langfengQ/verl-agent;

· HongbangYuan opened on Nov 25, 2025

Does verl-agent support use policy model itself as a tool executor?

Status: Open.

#201 In langfengQ/verl-agent;

· ImmortalSdm opened on Nov 20, 2025