List your company

Mathematical Reasoning on MATH 2021 (test)

78.3pass@1FedAvg-GRPO
54.27660.51366.7572.987Apr 14, 2026

Evaluation results

MethodLinks
FedAvg-GRPOModel=Qwen2-1.5B, Local steps (τ)=102026.0478.3
FedProx-GRPOModel=Qwen2-1.5B, Local steps (τ)=102026.0477.5
FedAvg-PubSwapModel=Qwen2-1.5B, Local steps (τ)=102026.0477
FedAvg-PubSwapModel=Qwen2-1.5B, Local steps (τ)=402026.0476.9
FedAvg-PubSwapModel=Qwen2-1.5B, Local steps (τ)=902026.0476.9
FedProx-GRPOModel=Qwen2-1.5B, Local steps (τ)=402026.0476.7
FedAvg-PubSwapModel=Qwen2-1.5B, Local steps (τ)=1202026.0476.6
FedProx-GRPOModel=Qwen2-1.5B, Local steps (τ)=902026.0476.5
FedAvg-GRPOModel=Qwen2-1.5B, Local steps (τ)=902026.0476.3
FedAvg-GRPOModel=Qwen2-1.5B, Local steps (τ)=402026.0476.1
FedAvg-GRPOModel=Qwen2-1.5B, Local steps (τ)=1202026.0475.9
FedProx-GRPOModel=Qwen2-1.5B, Local steps (τ)=1202026.0475.6
FedAvg-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=102026.0473.7
FedAvg-PubSwapModel=Qwen2.5-Math-1.5B, Local steps (τ)=402026.0473.5
FedAvg-PubSwapModel=Qwen2.5-Math-1.5B, Local steps (τ)=902026.0473.2
FedAvg-PubSwapModel=Qwen2.5-Math-1.5B, Local steps (τ)=102026.0473.1
FedAvg-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=402026.0473.1
FedProx-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=102026.0472.8
FedAvg-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=902026.0471.9
FedAvg-PubSwapModel=Qwen2.5-Math-1.5B, Local steps (τ)=1202026.0471.9
FedProx-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=902026.0471.5
FedProx-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=402026.0471.4
FedAvg-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=1202026.0471.1
FedProx-GRPOModel=Qwen2.5-Math-1.5B, Local steps (τ)=1202026.0470.6
Base modelModel=Qwen2.5-Math-1.5B, Local steps (τ)=102026.0458.4
Base modelModel=Qwen2.5-Math-1.5B, Local steps (τ)=402026.0458.4
Base modelModel=Qwen2.5-Math-1.5B, Local steps (τ)=902026.0458.4
Base modelModel=Qwen2.5-Math-1.5B, Local steps (τ)=1202026.0458.4
Base modelModel=Qwen2-1.5B, Local steps (τ)=102026.0455.2
Base modelModel=Qwen2-1.5B, Local steps (τ)=402026.0455.2
Base modelModel=Qwen2-1.5B, Local steps (τ)=902026.0455.2
Base modelModel=Qwen2-1.5B, Local steps (τ)=1202026.0455.2