Anwaarma's picture
Upload BertForSequenceClassification
f9e0d40 verified
|
raw
history blame
2.42 kB
metadata
base_model: Ammar-alhaj-ali/arabic-MARBERT-sentiment
metrics:
  - accuracy
tags:
  - generated_from_trainer
model-index:
  - name: unfortified_marbert2
    results: []

unfortified_marbert2

This model is a fine-tuned version of Ammar-alhaj-ali/arabic-MARBERT-sentiment on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1953
  • Accuracy: 0.91

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 0.3250 0.87
No log 0.1092 100 0.2651 0.91
No log 0.1638 150 0.4169 0.84
No log 0.2183 200 0.3473 0.89
No log 0.2729 250 0.3455 0.88
No log 0.3275 300 0.2863 0.9
No log 0.3821 350 0.2220 0.93
No log 0.4367 400 0.2382 0.92
No log 0.4913 450 0.3704 0.89
0.2918 0.5459 500 0.2641 0.92
0.2918 0.6004 550 0.2706 0.9
0.2918 0.6550 600 0.1759 0.93
0.2918 0.7096 650 0.3032 0.92
0.2918 0.7642 700 0.1880 0.91
0.2918 0.8188 750 0.2076 0.92
0.2918 0.8734 800 0.3046 0.91
0.2918 0.9279 850 0.1953 0.91

Framework versions

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