robust-marbert / README.md
Anwaarma's picture
Upload BertForSequenceClassification
92666b8 verified
metadata
base_model: Anwaarma/Improved-MARBERT-attempt2
metrics:
  - accuracy
tags:
  - generated_from_trainer
model-index:
  - name: robust-marbert
    results: []

robust-marbert

This model is a fine-tuned version of Anwaarma/Improved-MARBERT-attempt2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2362
  • Accuracy: 0.94

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 2e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Accuracy
No log 0.0546 50 0.2510 0.92
No log 0.1092 100 0.1780 0.94
No log 0.1638 150 0.3531 0.88
No log 0.2183 200 0.2775 0.94
No log 0.2729 250 0.2566 0.94
No log 0.3275 300 0.2247 0.94
No log 0.3821 350 0.1856 0.94
No log 0.4367 400 0.1221 0.96
No log 0.4913 450 0.3179 0.92
0.2513 0.5459 500 0.3608 0.9
0.2513 0.6004 550 0.1665 0.95
0.2513 0.6550 600 0.2186 0.93
0.2513 0.7096 650 0.2184 0.93
0.2513 0.7642 700 0.2175 0.93
0.2513 0.8188 750 0.2251 0.93
0.2513 0.8734 800 0.3068 0.92
0.2513 0.9279 850 0.1925 0.94
0.2513 0.9825 900 0.2141 0.93
0.2513 1.0371 950 0.2388 0.92
0.2118 1.0917 1000 0.3367 0.93
0.2118 1.1463 1050 0.2358 0.92
0.2118 1.2009 1100 0.3329 0.93
0.2118 1.2555 1150 0.2384 0.92
0.2118 1.3100 1200 0.3006 0.95
0.2118 1.3646 1250 0.2859 0.94
0.2118 1.4192 1300 0.2504 0.93
0.2118 1.4738 1350 0.2760 0.92
0.2118 1.5284 1400 0.2783 0.94
0.2118 1.5830 1450 0.2242 0.94
0.1485 1.6376 1500 0.2759 0.94
0.1485 1.6921 1550 0.2582 0.94
0.1485 1.7467 1600 0.3341 0.91
0.1485 1.8013 1650 0.3070 0.91
0.1485 1.8559 1700 0.1960 0.92
0.1485 1.9105 1750 0.2362 0.94

Framework versions

  • Transformers 4.42.2
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1