Tool Use on ToolMind
83.9Infer TimeOptimized Image (ART)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Optimized Image (ART)Backbone=Qwen3.5-0.8B, Hardware=NVIDIA A100, Inference Batch Size=2002026.06 | 83.9 | 1,158.6 | |
| BaselineBackbone=Qwen3.5-0.8B, Hardware=NVIDIA A100, Inference Batch Size=2002026.06 | 154.8 | — | |
| Fixed Seed ImageBackbone=Qwen3.5-0.8B, Hardware=NVIDIA A100, Inference Batch Size=2002026.06 | 180.3 | — | |
| LoRA (DAPO)Backbone=Qwen3.5-0.8B, Hardware=NVIDIA A100, Inference Batch Size=2002026.06 | 302.7 | 3,797.4 |