CodeBERTa

Reported on 8 benchmarks across 2 tasks

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Code8 results

Program SynthesisonManyTypes4TypeScript
Average Accuracy
59.81
best: 71.27 (CodeTIDAL5)
Program SynthesisonManyTypes4TypeScript
Average F1
56.71
best: 60.57 (GraphCodeBERT)
Program SynthesisonManyTypes4TypeScript
Average Precision
56.57
best: 60.06 (GraphCodeBERT)
Program SynthesisonManyTypes4TypeScript
Average Recall
56.85
best: 61.08 (GraphCodeBERT)
Type predictiononManyTypes4TypeScript
Average Accuracy
59.81
best: 71.27 (CodeTIDAL5)
Type predictiononManyTypes4TypeScript
Average F1
56.71
best: 60.57 (GraphCodeBERT)
Type predictiononManyTypes4TypeScript
Average Precision
56.57
best: 60.06 (GraphCodeBERT)
Type predictiononManyTypes4TypeScript
Average Recall
56.85
best: 61.08 (GraphCodeBERT)