Edit model card

robust-arabert

This model is a fine-tuned version of Anwaarma/Improved-Arabert-twitter-sentiment2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2072
  • 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 0.2677 0.91
No log 0.1092 100 0.2432 0.93
No log 0.1638 150 0.3328 0.91
No log 0.2183 200 0.2894 0.92
No log 0.2729 250 0.3008 0.91
No log 0.3275 300 0.1951 0.93
No log 0.3821 350 0.2176 0.94
No log 0.4367 400 0.2291 0.95
No log 0.4913 450 0.2678 0.93
0.2068 0.5459 500 0.2975 0.9
0.2068 0.6004 550 0.2245 0.92
0.2068 0.6550 600 0.2014 0.94
0.2068 0.7096 650 0.2189 0.93
0.2068 0.7642 700 0.2272 0.94
0.2068 0.8188 750 0.2148 0.92
0.2068 0.8734 800 0.2082 0.94
0.2068 0.9279 850 0.1869 0.93
0.2068 0.9825 900 0.2166 0.95
0.2068 1.0371 950 0.2095 0.94
0.1852 1.0917 1000 0.2948 0.92
0.1852 1.1463 1050 0.2586 0.93
0.1852 1.2009 1100 0.2669 0.93
0.1852 1.2555 1150 0.2635 0.94
0.1852 1.3100 1200 0.2600 0.94
0.1852 1.3646 1250 0.2290 0.95
0.1852 1.4192 1300 0.2248 0.95
0.1852 1.4738 1350 0.1840 0.96
0.1852 1.5284 1400 0.2366 0.95
0.1852 1.5830 1450 0.2322 0.95
0.1137 1.6376 1500 0.1797 0.96
0.1137 1.6921 1550 0.1662 0.96
0.1137 1.7467 1600 0.1762 0.96
0.1137 1.8013 1650 0.1508 0.96
0.1137 1.8559 1700 0.1996 0.95
0.1137 1.9105 1750 0.2072 0.94

Framework versions

  • Transformers 4.42.2
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
10
Safetensors
Model size
135M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for Anwaarma/robust-arabert