README.md

metadata

license: apache-2.0
base_model: LazarusNLP/IndoNanoT5-base
tags:
  - generated_from_trainer
datasets:
  - indonlg
metrics:
  - f1
model-index:
  - name: IndoNanoT5-base-TyDiQA
    results:
      - task:
          name: Sequence-to-sequence Language Modeling
          type: text2text-generation
        dataset:
          name: indonlg
          type: indonlg
          config: question_answering
          split: validation
          args: question_answering
        metrics:
          - name: F1
            type: f1
            value: 72.19688326266134

IndoNanoT5-base-TyDiQA

This model is a fine-tuned version of LazarusNLP/IndoNanoT5-base on the indonlg dataset. It achieves the following results on the evaluation set:

• Exact: 58.9474
• F1: 72.1969
• Total: 855
• Hasans Exact: 58.9474
• Hasans F1: 72.1969
• Hasans Total: 855
• Best Exact: 58.9474
• Best Exact Thresh: 0.0
• Best F1: 72.1969
• Best F1 Thresh: 0.0
• Loss: 0.1283

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Exact	F1	Total	Hasans Exact	Hasans F1	Hasans Total	Best Exact	Best Exact Thresh	Best F1	Best F1 Thresh	Validation Loss
1.9173	1.0	606	45.1327	63.8499	565	45.1327	63.8499	565	45.1327	0.0	63.8499	0.0	0.1147
0.1971	2.0	1212	50.4425	68.7240	565	50.4425	68.7240	565	50.4425	0.0	68.7240	0.0	0.1025
0.1475	3.0	1818	53.8053	71.0124	565	53.8053	71.0124	565	53.8053	0.0	71.0124	0.0	0.0992
0.1175	4.0	2424	53.6283	71.1353	565	53.6283	71.1353	565	53.6283	0.0	71.1353	0.0	0.1008
0.0814	5.0	3030	53.4513	71.0439	565	53.4513	71.0439	565	53.4513	0.0	71.0439	0.0	0.1040
0.0665	6.0	3636	54.1593	71.5788	565	54.1593	71.5788	565	54.1593	0.0	71.5788	0.0	0.1051
0.0555	7.0	4242	54.8673	72.4372	565	54.8673	72.4372	565	54.8673	0.0	72.4372	0.0	0.1137
0.0483	8.0	4848	56.2832	72.3749	565	56.2832	72.3749	565	56.2832	0.0	72.3749	0.0	0.1188
0.0416	9.0	5454	55.5752	72.2892	565	55.5752	72.2892	565	55.5752	0.0	72.2892	0.0	0.1154
0.031	10.0	6060	55.0442	71.8127	565	55.0442	71.8127	565	55.0442	0.0	71.8127	0.0	0.1312
0.0278	11.0	6666	55.7522	73.4756	565	55.7522	73.4756	565	55.7522	0.0	73.4756	0.0	0.1253
0.0257	12.0	7272	55.7522	73.0958	565	55.7522	73.0958	565	55.7522	0.0	73.0958	0.0	0.1292
0.023	13.0	7878	56.2832	73.3269	565	56.2832	73.3269	565	56.2832	0.0	73.3269	0.0	0.1271

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu118
• Datasets 2.16.1
• Tokenizers 0.15.1