# Fearao/RoBERTa_based_on_eastmoney_guba_comments

## Model description
This model is based on uer/roberta-base-finetuned-dianping-chinese, fine-tuned using comment data from Eastmoney stock bar, and I used the original tokenizer.
Thanks a lot to the authors of the model for all the help they gave me.

## How to use
You can use this model directly with a pipeline for text classification (take the case of RoBERTa_based_on_eastmoney_guba_comments):

```python
>>> from transformers import AutoModelForSequenceClassification,AutoTokenizer,pipeline
>>> model = AutoModelForSequenceClassification.from_pretrained('Fearao/RoBERTa_based_on_eastmoney_guba_comments')
>>> tokenizer = AutoTokenizer.from_pretrained('uer/roberta-base-finetuned-chinanews-chinese')
>>> text_classification = pipeline('sentiment-analysis', model=model, tokenizer=tokenizer)
>>> text_classification("又跳水了")
   [{'label': 'negative (stars 1, 2 and 3)', 'score': 0.9989427924156189}]
```

## Training data
Eastmoney stock bar comments datasets are used [Fearao/guba_eastmoney](https://huggingface.co/datasets/Fearao/guba_eastmoney)。

## Training procedure
Num examples = 7087 \
Num Epochs = 3 \
Instantaneous batch size per device = 8 \
Total train batch size (w. parallel, distributed & accumulation) = 8 \
Gradient Accumulation steps = 1 \
Total optimization steps = 2658 \
Number of trainable parameters = 102269186