SWE-Lego/qwen3_8b_glm5_swerebench_oraclesolved_oh_sdk_512_gbs64pbs1acc8_lr1e-4_epo3
Text Generation • 308k • Updated
AI & ML interests
We study code intelligence.
Recent Activity
View all activity
Papers
What Makes Interaction Trajectories Effective for Training Terminal Agents?
SWE-Lego: Pushing the Limits of Supervised Fine-tuning for Software Issue Resolving
SWE-Lego/SWE-Review-30B-A3B
Text Generation • 211k • Updated • 24
SWE-Lego/SWE-Review-8B
Text Generation • 308k • Updated • 27 • 1
SWE-Lego/Terminal-Lego-Qwen3-8B
Text Generation • 308k • Updated • 19 • 2
SWE-Lego/SWE-Lego-Verifier-30B-A3B
211k • Updated • 5
SWE-Lego/SWE-Lego-Verifier-8B
308k • Updated • 24
SWE-Lego/SWE-Lego-Qwen3-8B
308k • Updated • 488 • 5
SWE-Lego/SWE-Lego-Qwen3-32B
677k • Updated • 36 • 4