powerpuf-bot/wangchanberta-xet_ref-params

wangchanberta_dataxet_FAQ_chatbot_v3

This model is a fine-tuned version of Thammarak/wangchanBERTa-QA-thaiqa_squad on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0197

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
0.0301	0.53	100	0.0390
0.0377	1.07	200	0.0391
0.0338	1.6	300	0.0390
0.0153	2.14	400	0.0390
0.0152	2.67	500	0.0480
0.0166	3.21	600	0.0423
0.0206	3.74	700	0.0418
0.0285	4.28	800	0.0391
0.0522	4.81	900	0.0390
0.0339	5.35	1000	0.0197
0.0032	5.88	1100	0.0197
0.021	6.42	1200	0.0197
0.0073	6.95	1300	0.0197
0.016	7.49	1400	0.0197
0.0106	8.02	1500	0.0197
0.0081	8.56	1600	0.0197
0.0078	9.09	1700	0.0197
0.0001	9.63	1800	0.0197
0.0088	10.16	1900	0.0197
0.0187	10.7	2000	0.0198
0.0	11.23	2100	0.0198
0.0075	11.76	2200	0.0198
0.0145	12.3	2300	0.0197
0.0	12.83	2400	0.0197
0.0006	13.37	2500	0.0197
0.0069	13.9	2600	0.0197
0.0073	14.44	2700	0.0197
0.0002	14.97	2800	0.0197
0.0001	15.51	2900	0.0197
0.0234	16.04	3000	0.0197
0.0073	16.58	3100	0.0197
0.0001	17.11	3200	0.0197
0.008	17.65	3300	0.0197
0.0075	18.18	3400	0.0197
0.007	18.72	3500	0.0197
0.0	19.25	3600	0.0197
0.0208	19.79	3700	0.0197

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0