Qwen3-Coder-Next MoE models pruned for software engineering using REAP, gate-based, and activation-based expert importance.
tomsawyer
tomhu
·
AI & ML interests
None yet
Recent Activity
updated a model 4 days ago
tomhu/Qwen3-Coder-Next-SWE-Prune-Rebench-SFT-step480 published a model 4 days ago
tomhu/Qwen3-Coder-Next-SWE-Prune-Rebench-SFT-step480 updated a model 4 days ago
tomhu/Qwen3-Coder-Next-SWE-Prune-Rebench-SFT-final1086Organizations
None yet
models 52
tomhu/Qwen3-Coder-Next-SWE-Prune-Rebench-SFT-step480
26B • Updated • 10
tomhu/Qwen3-Coder-Next-SWE-Prune-Rebench-SFT-final1086
26B • Updated • 13
tomhu/Qwen3-Coder-Next-SWE-Prune-26B-A3B-Rebench-OPD
Updated • 3
tomhu/Qwen3-Coder-Next-SWE-Prune-26B-A3B-Rebench-GRPO
Updated • 15
tomhu/Qwen3-Coder-Next-SWE-Prune-GATE-70
Text Generation • 26B • Updated • 305
tomhu/Qwen3-Coder-Next-SWE-Prune-REAP-70
Text Generation • 26B • Updated • 186
tomhu/Qwen3-Coder-Next-SWE-Prune-ACTIVATION-70
Text Generation • 26B • Updated • 196
tomhu/RL4TG-Qwen3-8B-Offline-SFT-GRPO
Text Generation • 8B • Updated • 5
tomhu/RL4TG-Qwen3-8B-OPD-GRPO
Text Generation • 8B • Updated • 5
tomhu/RL4TG-Qwen3-8B-Direct-GRPO
Text Generation • 8B • Updated • 7