You can deploy this model and start earning money today!

Felixbrk/
robbert-2023-dutch-large-simple-score-text-only

safetensors

roberta

Views

No views yet

🚀 Felixbrk/robbert-2023-dutch-large-simple-score-text-only

This is a fine-tuned robbert-2023-dutch-large model for text-only regression on Dutch text.

📌 Model Summary

Base model: DTAI-KULeuven/robbert-2023-dutch-large
Task: Text Regression
Languages: Dutch (nl)
Heads: Re-initialized RobertaForSequenceClassification head for regression
Tokenizer: Pushed with model

📊 Training & Evaluation

Epoch	Training Loss	Validation Loss	RMSE	R²
1	0.017300	0.008135	0.0902	0.7814
2	0.009800	0.009247	0.0962	0.7516
3	0.007800	0.005978	0.0773	0.8394

✅ Test RMSE: 0.0773
✅ Test R²: 0.8394
✅ Head: transformer_y_quality_robbert2023 → multi-head aggregate

📂 How to use

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification

model = AutoModelForSequenceClassification.from_pretrained("Felixbrk/robbert-2023-dutch-large-simple-score-text-only")
tokenizer = AutoTokenizer.from_pretrained("Felixbrk/robbert-2023-dutch-large-simple-score-text-only")

inputs = tokenizer("Je voorbeeldzin hier.", return_tensors="pt")
outputs = model(**inputs)
print(outputs.logits)