scenario-kd-from-post-finetune-gold-silver-div-6-6000-data-smsa-model-haryoaw-sc

This model is a fine-tuned version of haryoaw/scenario-normal-finetune-clf-data-smsa-model-xlm-roberta-base on the smsa dataset. It achieves the following results on the evaluation set:

• Loss: 1.6482
• Accuracy: 0.8937
• F1: 0.8587

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6969

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.53	100	4.4646	0.7690	0.5226
No log	1.06	200	3.1571	0.8310	0.7511
No log	1.6	300	2.6080	0.8611	0.8199
No log	2.13	400	2.6827	0.8571	0.8156
3.6041	2.66	500	2.3719	0.8683	0.8262
3.6041	3.19	600	2.1546	0.8746	0.8407
3.6041	3.72	700	2.3281	0.8619	0.8241
3.6041	4.26	800	2.0646	0.8746	0.8333
3.6041	4.79	900	2.1933	0.8690	0.8271
1.4666	5.32	1000	2.1517	0.8762	0.8361
1.4666	5.85	1100	2.0496	0.8786	0.8367
1.4666	6.38	1200	2.0516	0.8738	0.8338
1.4666	6.91	1300	2.0529	0.8817	0.8433
1.4666	7.45	1400	1.8446	0.8841	0.8407
0.9173	7.98	1500	2.1299	0.8762	0.8295
0.9173	8.51	1600	1.9250	0.8921	0.8550
0.9173	9.04	1700	1.8133	0.8881	0.8384
0.9173	9.57	1800	1.8966	0.8873	0.8524
0.9173	10.11	1900	2.0855	0.8857	0.8501
0.6805	10.64	2000	1.8984	0.8881	0.8498
0.6805	11.17	2100	1.9069	0.8921	0.8504
0.6805	11.7	2200	1.9060	0.8865	0.8493
0.6805	12.23	2300	1.7894	0.8944	0.8600
0.6805	12.77	2400	1.9083	0.8825	0.8403
0.5396	13.3	2500	1.7880	0.8913	0.8471
0.5396	13.83	2600	1.9085	0.8849	0.8437
0.5396	14.36	2700	1.8370	0.8841	0.8405
0.5396	14.89	2800	1.8247	0.8817	0.8411
0.5396	15.43	2900	1.8115	0.8897	0.8458
0.495	15.96	3000	1.8270	0.8881	0.8561
0.495	16.49	3100	1.7943	0.8905	0.8470
0.495	17.02	3200	1.6322	0.8937	0.8550
0.495	17.55	3300	1.7213	0.8889	0.8490
0.495	18.09	3400	1.7923	0.8841	0.8344
0.4303	18.62	3500	1.7254	0.8905	0.8518
0.4303	19.15	3600	1.7596	0.8889	0.8481
0.4303	19.68	3700	1.6798	0.8865	0.8499
0.4303	20.21	3800	1.6482	0.8937	0.8587

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1
• Datasets 2.14.5
• Tokenizers 0.13.3