Code Efficiency Evaluation on EvalPerf
83.6DPSDeepSeek-Coder-6.7B-Instruct
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DeepSeek-Coder-6.7B-InstructTemperature=1.0, Samples per task=1002024.10 | 83.6 | 73.6 | 63.9 | |
| Llama-3.1-8B-InstructTemperature=1.0, Samples per task=1002024.10 | 80.9 | 64.3 | 52.1 | |
| CodeQwen1.5-7B-ChatTemperature=1.0, Samples per task=1002024.10 | 80.7 | 74.1 | 51.2 | |
| SelfCodeAlign-CQ-7BTemperature=1.0, Samples per task=1002024.10 | 79.9 | 65.2 | 54 | |
| Llama-3-8B-InstructTemperature=1.0, Samples per task=1002024.10 | 77 | 43.7 | 51.5 | |
| CodeQwen1.5-7B-OctoPackTemperature=1.0, Samples per task=1002024.10 | 74 | 49.1 | 26.9 |