These models are specialized for evaluating sentiment in financial news and market-related content. Unlike general-purpose sentiment models that often misclassify objective financial reporting as neutral, these models are fine-tuned to distinguish subtle positive, negative, and truly neutral signals in financial language.

Each language has its own dedicated model, trained and evaluated on high-quality financial datasets to capture language-specific nuances. This makes them particularly effective for applications such as market sentiment analysis, portfolio risk monitoring, trading signal generation, and financial news aggregation.

Chinese: https://huggingface.co/bardsai/finance-sentiment-zh-base
French: https://huggingface.co/bardsai/finance-sentiment-fr-base
German: https://huggingface.co/bardsai/finance-sentiment-de-base
Spanish: https://huggingface.co/bardsai/finance-sentiment-es-base
Japanese: https://huggingface.co/bardsai/finance-sentiment-ja-base
Polish: https://huggingface.co/bardsai/finance-sentiment-pl-base

How to use

You can use this model directly with a pipeline for sentiment-analysis:

from transformers import pipeline

nlp = pipeline("sentiment-analysis", model="bardsai/finance-sentiment-fr-base")
nlp("Le chiffre d'affaires net a augmenté de 30 % pour atteindre 36 millions d'euros.")

[{'label': 'positive', 'score': 0.9999314546585083}]

Performance

Chinese

Polish

French

German

Japanese

Spanish