### deberta-v3-large finetuned on a mixture of three sentiment classification datasets

#### train set
sst2<br>
financial_phrasebank<br>
imdb<br>

#### test set
sst2<br>
imdb<br>

#### train args<br>
max_input_length: 256<br>
epoch: 3<br>
optimizer: AdamW<br>
lr: 1e-5<br>
fp16: True<br>

#### performance<br>
accuracy on mixture of sst2 and imdb test set: 95.3%<br>