twhin-bert-base

This model is a fine-tuned version of Twitter/twhin-bert-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6341
• F1: 0.3077

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-07
• 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.6867	1.0	189	0.6817	0.1026
0.684	2.0	378	0.6746	0.0571
0.675	3.0	567	0.6649	0.0
0.6642	4.0	756	0.6577	0.0
0.6653	5.0	945	0.6542	0.0
0.6649	6.0	1134	0.6479	0.0
0.6648	7.0	1323	0.6460	0.0
0.6511	8.0	1512	0.6387	0.0
0.6535	9.0	1701	0.6332	0.0
0.6544	10.0	1890	0.6261	0.0
0.6382	11.0	2079	0.6154	0.0
0.6315	12.0	2268	0.6051	0.0
0.6231	13.0	2457	0.5957	0.2326
0.603	14.0	2646	0.5858	0.2326
0.6034	15.0	2835	0.5771	0.2553
0.5938	16.0	3024	0.5694	0.2308
0.5884	17.0	3213	0.5642	0.3103
0.5763	18.0	3402	0.5611	0.3103
0.5675	19.0	3591	0.5641	0.2857
0.5672	20.0	3780	0.5598	0.3000
0.5674	21.0	3969	0.5579	0.2857
0.5479	22.0	4158	0.5642	0.3125
0.5621	23.0	4347	0.5688	0.2903
0.5516	24.0	4536	0.5685	0.3077
0.5597	25.0	4725	0.5713	0.3077
0.5418	26.0	4914	0.5761	0.3077
0.5477	27.0	5103	0.5752	0.3030
0.535	28.0	5292	0.5876	0.3077
0.5544	29.0	5481	0.5841	0.3030
0.5238	30.0	5670	0.5855	0.3030
0.5375	31.0	5859	0.5894	0.3030
0.5092	32.0	6048	0.5985	0.3077
0.5262	33.0	6237	0.5988	0.3077
0.5418	34.0	6426	0.6038	0.3077
0.531	35.0	6615	0.6087	0.3077
0.5627	36.0	6804	0.6064	0.3077
0.545	37.0	6993	0.6110	0.3077
0.5105	38.0	7182	0.6134	0.3077
0.5471	39.0	7371	0.6111	0.3077
0.5114	40.0	7560	0.6212	0.3077
0.5411	41.0	7749	0.6159	0.3077
0.5304	42.0	7938	0.6213	0.3077
0.5146	43.0	8127	0.6276	0.3077
0.5223	44.0	8316	0.6301	0.3077
0.5345	45.0	8505	0.6281	0.3077
0.5368	46.0	8694	0.6284	0.3077
0.516	47.0	8883	0.6320	0.3077
0.5241	48.0	9072	0.6339	0.3077
0.5267	49.0	9261	0.6342	0.3077
0.5478	50.0	9450	0.6341	0.3077

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.2
• Datasets 2.12.0
• Tokenizers 0.13.3