Bullish-bearish sentiment labels
The model card describes Bullish and Bearish user labels, with label 0 as bearish and label 1 as bullish.
Open Source Model Profile · zhayunduo
roberta-base-stocktwits-finetuned
roberta-base-stocktwits-finetuned is a zhayunduo RoBERTa classifier for Bullish-versus-Bearish stock sentiment. Its card documents fine-tuning on 3.2M Stocktwits comments.
- Publisher
- zhayunduo
- Task
- text-classification
- Model type
- roberta
- License
- apache-2.0
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
roberta-base-stocktwits-finetuned is published by zhayunduo as a text-classification model for stock-comment sentiment. The captured configuration identifies RobertaForSequenceClassification with model type roberta. According to the model card, it is fine-tuned from roberta-base on 3,200,000 Stocktwits comments labeled Bullish or Bearish.
Recorded capabilities
3.2M-comment Stocktwits fine-tune
According to the model card, the roberta-base model was fine-tuned on 3,200,000 Stocktwits comments, with reported batch size 32, learning rate 2e-5, and four-epoch loss and accuracy figures.
Documented preprocessing and loading
The model card documents a text preprocessing routine for URLs, cashtags, hashtags, mentions, and emoji, plus RobertaTokenizer and RobertaForSequenceClassification loading calls.
Use cases in the source record
- Bullish-versus-Bearish sentiment scoring for stock-forum style comments.
- Finance text-classification experiments on English Stocktwits-style text that can apply the documented preprocessing routine.
Limitations and unknowns
- No parameter count was extracted from this record.
- No evaluation results beyond the publisher-reported training loss and accuracy table were extracted.
- Provider state is historical snapshot data and should be refreshed before being presented as current availability.
- Training data and accuracy figures come from the publisher model card and have not been independently verified by Ethen.
Source and provenance
Source: zhayunduo/roberta-base-stocktwits-finetuned
Captured: Unknown. Processed: 2026-09-07T19:35:01.757727+00:00.
Sentiment Inferencing model for stock related commments A project by NUS ISS students Frank Cao, Gerong Zhang, Jiaqi Yao, Sikai Ni, Yunduo Zhang Description This model is fine tuned with roberta-base model on 3200000 comments from stocktwits, with the user labeled tags 'Bullish' or 'Bearish' try something that the individual investors may say on the investment forum on the inference API, for example, try 'red' and 'green'. code on github Training information batch size 32 learning rate 2e-5 Train loss Validation loss Validation accuracy epoch1 0.3495 0.2956 0.8679 epoch2 0.2717 0.2235 0.9021 epoch3 0.2360 0.1875 0.9210 epoch4 0.2106…
F001F002F003F004F005F006F007F008F009F010F011F012F013F014