lora-roberta-large-no-ed

This model is a fine-tuned version of roberta-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6852
• Accuracy: 0.7581
• Prec: 0.6440
• Recall: 0.6039
• F1: 0.6206
• B Acc: 0.6039
• Micro F1: 0.7581
• Prec Joy: 0.7167
• Recall Joy: 0.7642
• F1 Joy: 0.7397
• Prec Anger: 0.6203
• Recall Anger: 0.6289
• F1 Anger: 0.6246
• Prec Disgust: 0.4767
• Recall Disgust: 0.3849
• F1 Disgust: 0.4259
• Prec Fear: 0.6752
• Recall Fear: 0.5813
• F1 Fear: 0.6247
• Prec Neutral: 0.8418
• Recall Neutral: 0.8525
• F1 Neutral: 0.8471
• Prec Sadness: 0.6674
• Recall Sadness: 0.6614
• F1 Sadness: 0.6644
• Prec Surprise: 0.5101
• Recall Surprise: 0.3542
• F1 Surprise: 0.4181

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: 0.001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Prec	Recall	F1	B Acc	Micro F1	Prec Joy	Recall Joy	F1 Joy	Prec Anger	Recall Anger	F1 Anger	Prec Disgust	Recall Disgust	F1 Disgust	Prec Fear	Recall Fear	F1 Fear	Prec Neutral	Recall Neutral	F1 Neutral	Prec Sadness	Recall Sadness	F1 Sadness	Prec Surprise	Recall Surprise	F1 Surprise
0.7938	1.0	1465	0.7589	0.7257	0.6233	0.4993	0.5433	0.4993	0.7257	0.7259	0.6828	0.7037	0.6223	0.4082	0.4930	0.5359	0.2657	0.3552	0.5925	0.5110	0.5487	0.7564	0.9097	0.8260	0.7150	0.4865	0.5790	0.4151	0.2315	0.2972
0.7546	2.0	2930	0.7482	0.7243	0.6272	0.5499	0.5735	0.5499	0.7243	0.6028	0.8315	0.6989	0.5325	0.5802	0.5553	0.5135	0.2782	0.3609	0.6619	0.5388	0.5940	0.8498	0.8045	0.8265	0.74	0.5294	0.6172	0.4902	0.2864	0.3616
0.7289	3.0	4395	0.7293	0.7321	0.6234	0.5839	0.5984	0.5839	0.7321	0.6491	0.7901	0.7127	0.6129	0.5271	0.5668	0.4413	0.4561	0.4486	0.6974	0.5198	0.5956	0.8364	0.8146	0.8254	0.6406	0.6423	0.6414	0.4862	0.3376	0.3985
0.7076	4.0	5860	0.6898	0.7466	0.6572	0.5649	0.5972	0.5649	0.7466	0.7573	0.6911	0.7227	0.5110	0.6565	0.5747	0.4868	0.3096	0.3785	0.8139	0.4802	0.6041	0.8125	0.8772	0.8436	0.6939	0.5946	0.6404	0.5253	0.3453	0.4167
0.6925	5.0	7325	0.7039	0.7403	0.6121	0.5916	0.5972	0.5916	0.7403	0.6933	0.7525	0.7217	0.5234	0.6372	0.5747	0.3630	0.4100	0.3851	0.6121	0.5798	0.5955	0.8446	0.8363	0.8404	0.7512	0.5713	0.6490	0.4973	0.3542	0.4137
0.6841	6.0	8790	0.6704	0.7516	0.6607	0.5820	0.6076	0.5820	0.7516	0.7158	0.7536	0.7342	0.6577	0.4856	0.5587	0.4195	0.5502	0.4760	0.8476	0.4641	0.5998	0.8120	0.8784	0.8439	0.6971	0.6184	0.6554	0.4756	0.3235	0.3851
0.6715	7.0	10255	0.6919	0.7412	0.6246	0.6180	0.6112	0.6180	0.7412	0.7020	0.7642	0.7318	0.5513	0.6034	0.5762	0.3682	0.5962	0.4553	0.8024	0.4817	0.6020	0.8602	0.8191	0.8391	0.6611	0.6481	0.6545	0.4267	0.4130	0.4198
0.6562	8.0	11720	0.7245	0.7325	0.5985	0.6129	0.6014	0.6129	0.7325	0.6499	0.8167	0.7238	0.5320	0.6211	0.5731	0.3779	0.4728	0.4201	0.5704	0.6047	0.5871	0.8771	0.7863	0.8292	0.7431	0.6010	0.6645	0.4391	0.3875	0.4117
0.6426	9.0	13185	0.6683	0.7510	0.6304	0.6109	0.6175	0.6109	0.7510	0.7216	0.7506	0.7358	0.5768	0.6001	0.5882	0.3908	0.4603	0.4227	0.7469	0.5359	0.6240	0.8458	0.8458	0.8458	0.6966	0.6412	0.6678	0.4347	0.4425	0.4385
0.6278	10.0	14650	0.6661	0.7545	0.6427	0.5968	0.6142	0.5968	0.7545	0.7531	0.712	0.7320	0.6346	0.5476	0.5879	0.4574	0.4268	0.4416	0.7220	0.5476	0.6228	0.8304	0.8692	0.8494	0.5931	0.7276	0.6535	0.5084	0.3465	0.4122
0.6218	11.0	16115	0.6714	0.7507	0.6478	0.5958	0.6143	0.5958	0.7507	0.6878	0.7864	0.7338	0.6796	0.4950	0.5728	0.4181	0.4916	0.4519	0.7635	0.4963	0.6016	0.8324	0.8512	0.8417	0.6816	0.6524	0.6667	0.4719	0.3977	0.4316
0.6077	12.0	17580	0.6649	0.7543	0.6216	0.6171	0.6187	0.6171	0.7543	0.7496	0.7249	0.7371	0.6055	0.6095	0.6075	0.4449	0.4142	0.4290	0.6194	0.6076	0.6135	0.8426	0.8568	0.8497	0.6894	0.6386	0.6630	0.4	0.4680	0.4313
0.5868	13.0	19045	0.6680	0.7584	0.6473	0.6026	0.6224	0.6026	0.7584	0.7192	0.7522	0.7354	0.6442	0.5658	0.6025	0.4398	0.4435	0.4417	0.7127	0.5666	0.6313	0.8293	0.8711	0.8497	0.7187	0.6174	0.6642	0.4673	0.4015	0.4319
0.5747	14.0	20510	0.6692	0.7551	0.6293	0.6049	0.6155	0.6049	0.7551	0.7114	0.7621	0.7359	0.5985	0.6167	0.6075	0.4461	0.3808	0.4108	0.6088	0.6061	0.6075	0.8444	0.8522	0.8483	0.7124	0.6222	0.6642	0.4835	0.3939	0.4341
0.5632	15.0	21975	0.6763	0.7551	0.6390	0.6104	0.6185	0.6104	0.7551	0.6978	0.7812	0.7371	0.6381	0.5774	0.6063	0.4179	0.5272	0.4662	0.6260	0.5710	0.5972	0.8432	0.8479	0.8455	0.6950	0.6460	0.6696	0.5551	0.3223	0.4078
0.546	16.0	23440	0.6880	0.7537	0.6365	0.6089	0.6205	0.6089	0.7537	0.6906	0.7878	0.7360	0.6121	0.625	0.6185	0.4564	0.3828	0.4164	0.6587	0.6076	0.6321	0.8493	0.8350	0.8421	0.6999	0.6428	0.6702	0.4885	0.3811	0.4282
0.5354	17.0	24905	0.6823	0.7545	0.6399	0.6097	0.6222	0.6097	0.7545	0.6972	0.7828	0.7375	0.6131	0.6355	0.6241	0.4916	0.3682	0.4211	0.6979	0.5783	0.6325	0.8525	0.8357	0.8440	0.6820	0.6455	0.6632	0.4447	0.4220	0.4331
0.5103	18.0	26370	0.6852	0.7581	0.6440	0.6039	0.6206	0.6039	0.7581	0.7167	0.7642	0.7397	0.6203	0.6289	0.6246	0.4767	0.3849	0.4259	0.6752	0.5813	0.6247	0.8418	0.8525	0.8471	0.6674	0.6614	0.6644	0.5101	0.3542	0.4181
0.4972	19.0	27835	0.6948	0.7535	0.6350	0.6039	0.6162	0.6039	0.7535	0.7038	0.7715	0.7361	0.5989	0.6515	0.6241	0.4658	0.3703	0.4126	0.6739	0.5871	0.6275	0.8495	0.8381	0.8438	0.6631	0.6571	0.6601	0.4902	0.3517	0.4095
0.4801	20.0	29300	0.6945	0.7549	0.6320	0.6106	0.6199	0.6106	0.7549	0.7138	0.7631	0.7376	0.6130	0.6316	0.6222	0.4396	0.4038	0.4209	0.6623	0.5886	0.6233	0.8481	0.8428	0.8455	0.6626	0.6619	0.6622	0.4846	0.3824	0.4274

Framework versions

• Transformers 4.32.0.dev0
• Pytorch 2.0.1
• Datasets 2.12.0
• Tokenizers 0.11.0