metadata
language:
- multilingual
tags:
- xlm-roberta
- sentiment-analysis
- text-classification
metrics:
- f1
pipeline_tag: text-classification
CMPE 346 - Assignment 02: Multilingual Sentiment Classification
Model Details
- Base Model: xlm-roberta-large
- Task: Binary sentiment classification
Training
- Epochs: 3
- Learning Rate: 1e-5
- Batch Size: 16
- Max Length: 512
- Weight Decay: 0.01
- FP16: True
- GPU: A100
Results
| Epoch | Training Loss | Validation Loss | F1 |
|---|---|---|---|
| 1 | 0.4330 | 0.2066 | 0.9244 |
| 2 | 0.3413 | 0.2147 | 0.9283 |
| 3 | 0.2684 | 0.2398 | 0.9279 |
Final Validation F1 Score: 0.9282