Edit model card

AraElectra-QA-EgyLaw-Squad

This model is a fine-tuned version of ZeyadAhmed/AraElectra-Arabic-SQuADv2-QA on the Egyption Law Squad dataset. It achieves the following results on the evaluation set:

  • Loss: 4.0008

Model description

Araelctra model for Question Answering Task in Egyption law specially Personal Status Law (ู‚ุงู†ูˆู† ุงู„ุงุญูˆุงู„ ุงู„ุดุฎุตูŠุฉ)

this Model was created for a Graduation Project in Computers and Artificial intellgence at Helwan University under supervisation from Dr.Ensaf Hossen

About Team

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 3e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 20

Training results

Training Loss Epoch Step Validation Loss
No log 1.0 89 3.2762
3.9126 2.0 178 3.0150
2.5679 3.0 267 2.8324
1.7576 4.0 356 2.9655
1.1946 5.0 445 3.1643
0.8854 6.0 534 3.0778
0.7387 7.0 623 3.2814
0.6067 8.0 712 3.2391
0.5201 9.0 801 3.3717
0.5201 10.0 890 3.5752
0.4377 11.0 979 3.6657
0.4025 12.0 1068 3.8178
0.4077 13.0 1157 3.7404
0.3355 14.0 1246 3.7566
0.3627 15.0 1335 3.9959
0.3376 16.0 1424 3.8815
0.3676 17.0 1513 3.9674
0.3165 18.0 1602 3.9368
0.3165 19.0 1691 3.9836
0.3148 20.0 1780 4.0008

Framework versions

  • Transformers 4.38.1
  • Pytorch 2.1.2
  • Datasets 2.1.0
  • Tokenizers 0.15.2
Downloads last month
4
Safetensors
Model size
135M params
Tensor type
F32
ยท
Inference API
This model can be loaded on Inference API (serverless).

Finetuned from

Dataset used to train BoodyAhmedHamdy/AraElectra-QA-EgyLaw-Squad