floriandebaene/gysbert_EmDComF_emotion

gysbert_historical_fmp2_ogtok_output_emotion_primary

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0069
• Accuracy: 0.7123
• F1: 0.5402
• Precision: 0.6014
• Recall: 0.5229

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-06
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 200
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.6286	0.3030	100	1.4354	0.5622	0.1200	0.0937	0.1667
1.2292	0.6061	200	1.0157	0.6065	0.2132	0.2454	0.2536
1.0055	0.9091	300	0.9542	0.6491	0.2868	0.4298	0.3134
0.9543	1.2121	400	0.9205	0.6610	0.3287	0.3193	0.3487
0.8554	1.5152	500	0.8846	0.6797	0.3406	0.3563	0.3626
0.8496	1.8182	600	0.8515	0.6780	0.3376	0.3400	0.3606
0.7995	2.1212	700	0.8363	0.6797	0.3599	0.3510	0.3806
0.7274	2.4242	800	0.8196	0.6814	0.3748	0.5206	0.3849
0.7402	2.7273	900	0.8130	0.6848	0.4108	0.4717	0.4126
0.7739	3.0303	1000	0.8050	0.6951	0.4295	0.4618	0.4362
0.6437	3.3333	1100	0.8085	0.6951	0.4475	0.5600	0.4309
0.653	3.6364	1200	0.8192	0.6917	0.4540	0.5471	0.4510
0.6566	3.9394	1300	0.8089	0.6917	0.4642	0.5545	0.4568
0.5754	4.2424	1400	0.8122	0.6968	0.4751	0.5435	0.4710
0.5799	4.5455	1500	0.8404	0.6882	0.4636	0.5242	0.4669
0.5898	4.8485	1600	0.8306	0.7104	0.5037	0.5878	0.4894
0.4972	5.1515	1700	0.8506	0.6985	0.4996	0.5380	0.4953
0.5047	5.4545	1800	0.8497	0.7002	0.4994	0.5374	0.4930
0.4803	5.7576	1900	0.8772	0.6968	0.4752	0.5421	0.4634
0.4932	6.0606	2000	0.8752	0.6899	0.4817	0.5110	0.4777
0.4104	6.3636	2100	0.8885	0.6985	0.4884	0.5257	0.4849
0.4255	6.6667	2200	0.9097	0.6882	0.4862	0.5055	0.4855
0.4752	6.9697	2300	0.9153	0.7019	0.4990	0.5266	0.4968
0.3682	7.2727	2400	0.9370	0.6865	0.4866	0.5051	0.4934
0.3738	7.5758	2500	0.9578	0.6882	0.4779	0.5069	0.4764
0.3342	7.8788	2600	0.9608	0.6917	0.5202	0.5305	0.5221
0.3324	8.1818	2700	0.9839	0.6797	0.5076	0.5069	0.5174
0.3247	8.4848	2800	0.9890	0.6814	0.4860	0.5048	0.4888
0.3101	8.7879	2900	1.0068	0.6814	0.4793	0.5039	0.4832
0.2835	9.0909	3000	1.0158	0.6882	0.5092	0.5262	0.5085
0.2792	9.3939	3100	1.0497	0.6831	0.4888	0.5233	0.4885
0.2426	9.6970	3200	1.0563	0.6882	0.5163	0.5227	0.5182
0.2636	10.0	3300	1.0859	0.6882	0.5219	0.5317	0.5278
0.2281	10.3030	3400	1.1076	0.6831	0.5040	0.5189	0.5073
0.238	10.6061	3500	1.1152	0.6865	0.4896	0.5059	0.4945
0.1861	10.9091	3600	1.1460	0.6848	0.4907	0.5062	0.4928

Framework versions

• Transformers 4.40.2
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.19.1