Edit model card

scam-alert-mobile-bert

This model is a fine-tuned version of google/mobilebert-uncased on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.7097
  • Accuracy: 0.9880
  • F1: 0.9880

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: 8
  • eval_batch_size: 2
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 6

Training results

Training Loss Epoch Step Validation Loss Accuracy F1
No log 0.1577 100 0.4729 0.9223 0.9145
No log 0.3155 200 2.1621 0.9801 0.9803
No log 0.4732 300 0.8327 0.9900 0.9900
No log 0.6309 400 3.3648 0.9900 0.9900
No log 0.7886 500 0.8376 0.9861 0.9861
No log 0.9464 600 0.7630 0.9861 0.9861
No log 1.1041 700 0.6559 0.9861 0.9861
No log 1.2618 800 2.2440 0.9880 0.9880
No log 1.4196 900 2.4358 0.9900 0.9900
No log 1.5773 1000 1.9655 0.9861 0.9859
No log 1.7350 1100 1.8927 0.9880 0.9880
No log 1.8927 1200 1.3919 0.9880 0.9880
No log 2.0505 1300 0.9143 0.9861 0.9860
No log 2.2082 1400 0.1891 0.9861 0.9859
No log 2.3659 1500 0.0815 0.9861 0.9861
No log 2.5237 1600 0.0853 0.9880 0.9880
No log 2.6814 1700 0.2719 0.9861 0.9860
No log 2.8391 1800 0.2175 0.9900 0.9900
No log 2.9968 1900 0.5407 0.9880 0.9880
No log 3.1546 2000 0.8695 0.9880 0.9880
No log 3.3123 2100 0.1031 0.9880 0.9880
No log 3.4700 2200 1.1922 0.9900 0.9900
No log 3.6278 2300 0.4830 0.9880 0.9880
No log 3.7855 2400 1.4562 0.9880 0.9880
No log 3.9432 2500 1.8929 0.9900 0.9900
2789.4062 4.1009 2600 0.6560 0.9880 0.9880
2789.4062 4.2587 2700 0.1473 0.9841 0.9842
2789.4062 4.4164 2800 0.3488 0.9880 0.9880
2789.4062 4.5741 2900 0.2347 0.9880 0.9880
2789.4062 4.7319 3000 0.7488 0.9900 0.9900
2789.4062 4.8896 3100 0.5055 0.9880 0.9880
2789.4062 5.0473 3200 0.8339 0.9900 0.9900
2789.4062 5.2050 3300 0.5382 0.9880 0.9880
2789.4062 5.3628 3400 0.6095 0.9880 0.9880
2789.4062 5.5205 3500 0.7142 0.9880 0.9880
2789.4062 5.6782 3600 0.6855 0.9880 0.9880
2789.4062 5.8360 3700 0.7152 0.9880 0.9880
2789.4062 5.9937 3800 0.7097 0.9880 0.9880

Framework versions

  • Transformers 4.41.1
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.2
  • Tokenizers 0.19.1
Downloads last month
8
Safetensors
Model size
24.6M 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 JeswinMS4/scam-alert-mobile-bert

Finetuned
(24)
this model