bert-medium-tiny / summary.txt
takedarn's picture
Upload fine-tuned model
3aa50a2 verified
=======================================
Model Training Summary
=======================================
Model: prajjwal1/bert-tiny
Task: wnli
Training Date: Sat Sep 6 02:53:11 PM JST 2025
Hyperparameters:
- Epochs: 5
- Batch Size: 32
- Learning Rate: 2e-5
- Warmup Steps: 200
- Weight Decay: 0.01
- Max Sequence Length: 128
Training Status: SUCCESS
Job: 15/15
Model Directory: /home/ubuntu/master-research/not-trust-quivalent/models/wnli/prajjwal1-bert-tiny
Script Path: /home/ubuntu/master-research/not-trust-quivalent/src/finetune/train.py
Output Directory: /home/ubuntu/master-research/not-trust-quivalent/models
=======================================