Irony Detection on TweetEval irony (test)
84.18AccuracyQ-RUN
Evaluation results
| Method | Links | |
|---|---|---|
| Q-RUNBackbone=Mistral-7b, Evaluation Protocol=Fine-tuning, Rank (r)=8, #Params (%)=0.0945, Time (min)=3.742025.12 | 84.18 | |
| DoRABackbone=Mistral-7b, Evaluation Protocol=Fine-tuning, Rank (r)=8, #Params (%)=0.0985, Time (min)=3.882025.12 | 83.16 | |
| Q-RUNBackbone=Mistral-7b, Evaluation Protocol=Fine-tuning, Rank (r)=4, #Params (%)=0.0476, Time (min)=3.732025.12 | 82.65 | |
| LoRABackbone=Mistral-7b, Evaluation Protocol=Fine-tuning, Rank (r)=8, #Params (%)=0.0939, Time (min)=3.612025.12 | 82.02 | |
| LoRABackbone=Mistral-7b, Evaluation Protocol=Fine-tuning, Rank (r)=4, #Params (%)=0.0470, Time (min)=3.592025.12 | 81.76 | |
| DoRABackbone=Mistral-7b, Evaluation Protocol=Fine-tuning, Rank (r)=4, #Params (%)=0.0515, Time (min)=3.692025.12 | 80.61 | |
| BaselineBackbone=Mistral-7b, Evaluation Protocol=Zero-shot inference, #Params (%)=100.02025.12 | 62.12 |