SentimentT2_GPT2

This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0308
• Accuracy: 0.8644
• F1: 0.8685
• Auc Roc: 0.9297
• Log Loss: 1.0307

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: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Auc Roc	Log Loss
1.1785	0.15	500	0.7334	0.8346	0.8400	0.9144	0.7334
1.1409	0.31	1000	0.8797	0.8520	0.8649	0.9269	0.8796
1.0906	0.46	1500	0.7869	0.8744	0.8805	0.9394	0.7869
1.0163	0.62	2000	0.8381	0.8706	0.8771	0.9366	0.8381
1.0602	0.77	2500	0.9904	0.8458	0.8616	0.9253	0.9904
1.1456	0.93	3000	0.8833	0.8483	0.8452	0.9275	0.8832
0.9662	1.08	3500	0.9737	0.8507	0.8618	0.9354	0.9737
0.8496	1.24	4000	0.9361	0.8619	0.8680	0.9351	0.9361
0.8571	1.39	4500	0.8660	0.8619	0.8702	0.9346	0.8660
0.7506	1.55	5000	0.9359	0.8507	0.8558	0.9316	0.9359
0.8236	1.7	5500	1.1721	0.8184	0.8433	0.9229	1.1721
0.6897	1.85	6000	0.9876	0.8532	0.8547	0.9318	0.9876
0.6699	2.01	6500	0.8947	0.8570	0.8671	0.9323	0.8946
0.6137	2.16	7000	0.9318	0.8557	0.8661	0.9344	0.9318
0.4646	2.32	7500	0.9943	0.8595	0.8660	0.9312	0.9944
0.7042	2.47	8000	0.9150	0.8657	0.8714	0.9345	0.9150
0.4079	2.63	8500	1.0215	0.8657	0.8750	0.9312	1.0214
0.4646	2.78	9000	0.9809	0.8619	0.8714	0.9310	0.9809
0.4707	2.94	9500	1.0151	0.8644	0.8719	0.9279	1.0150
0.5005	3.09	10000	1.0748	0.8607	0.8651	0.9289	1.0747
0.3817	3.24	10500	0.8819	0.8781	0.8858	0.9299	0.8818
0.279	3.4	11000	1.0542	0.8607	0.8627	0.9302	1.0541
0.3527	3.55	11500	1.0148	0.8607	0.8637	0.9312	1.0147
0.3873	3.71	12000	1.0421	0.8619	0.8648	0.9294	1.0420
0.3552	3.86	12500	1.0308	0.8644	0.8685	0.9297	1.0307

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1