Bencode92/tradepulse-finbert-correlations
Description
Fine-tuned FinBERT model for financial correlations analysis in TradePulse.
Task: Correlations Classification
Target Column: correlations
Multi-Label: Yes (61 labels)
Performance
Last training: 2025-07-30 12:03
Dataset: base_reference.csv (708 samples)
| Metric | Value |
|---|---|
| Loss | 0.1960 |
| Subset Accuracy | 0.0000 |
| F1 Score | 0.0000 |
| F1 Micro | 0.0000 |
| F1 Macro | 0.0000 |
| Hamming Score | 0.9799 |
| Precision | 0.0000 |
| Recall | 0.0000 |
Training Details
- Base Model: Bencode92/tradepulse-finbert-correlations
- Training Mode: Incremental
- Epochs: 2
- Learning Rate: 1e-05
- Batch Size: 4
- Class Balancing: None
- Problem Type: Multi-Label Classification
Usage
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
tokenizer = AutoTokenizer.from_pretrained("Bencode92/tradepulse-finbert-correlations")
model = AutoModelForSequenceClassification.from_pretrained("Bencode92/tradepulse-finbert-correlations")
# Example prediction
text = "Apple reported strong quarterly earnings beating expectations"
inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
outputs = model(**inputs)
# Multi-label: apply sigmoid and threshold
predictions = torch.sigmoid(outputs.logits).squeeze() > 0.5
Model Card Authors
- TradePulse ML Team
- Auto-generated on 2025-07-30 12:03:53