model_broadclass_onSet2.1

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1459
• 0 Precision: 0.9630
• 0 Recall: 1.0
• 0 F1-score: 0.9811
• 0 Support: 26
• 1 Precision: 1.0
• 1 Recall: 0.9231
• 1 F1-score: 0.9600
• 1 Support: 39
• 2 Precision: 1.0
• 2 Recall: 1.0
• 2 F1-score: 1.0
• 2 Support: 19
• 3 Precision: 0.8667
• 3 Recall: 1.0
• 3 F1-score: 0.9286
• 3 Support: 13
• Accuracy: 0.9691
• Macro avg Precision: 0.9574
• Macro avg Recall: 0.9808
• Macro avg F1-score: 0.9674
• Macro avg Support: 97
• Weighted avg Precision: 0.9722
• Weighted avg Recall: 0.9691
• Weighted avg F1-score: 0.9693
• Weighted avg Support: 97
• Wer: 0.1293
• Mtrix: [[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 1, 36, 0, 2], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 200
• num_epochs: 80
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	0 Precision	0 Recall	0 F1-score	0 Support	1 Precision	1 Recall	1 F1-score	1 Support	2 Precision	2 Recall	2 F1-score	2 Support	3 Precision	3 Recall	3 F1-score	3 Support	Accuracy	Macro avg Precision	Macro avg Recall	Macro avg F1-score	Macro avg Support	Weighted avg Precision	Weighted avg Recall	Weighted avg F1-score	Weighted avg Support	Wer	Mtrix
2.3399	4.16	100	2.1769	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
2.3152	8.33	200	2.1458	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
1.9859	12.49	300	1.9172	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
1.7126	16.65	400	1.6954	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
1.6833	20.82	500	1.7553	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
1.5318	24.98	600	1.5921	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
1.5868	29.16	700	1.5517	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
1.5577	33.33	800	1.5089	0.2680	1.0	0.4228	26	0.0	0.0	0.0	39	0.0	0.0	0.0	19	0.0	0.0	0.0	13	0.2680	0.0670	0.25	0.1057	97	0.0718	0.2680	0.1133	97	0.9869	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 39, 0, 0, 0], [2, 19, 0, 0, 0], [3, 13, 0, 0, 0]]
1.2201	37.49	900	1.1567	0.4643	1.0	0.6341	26	1.0	0.4872	0.6552	39	1.0	0.5263	0.6897	19	1.0	0.9231	0.9600	13	0.6907	0.8661	0.7341	0.7347	97	0.8564	0.6907	0.6971	97	0.9485	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 20, 19, 0, 0], [2, 9, 0, 10, 0], [3, 1, 0, 0, 12]]
0.9692	41.65	1000	1.0489	0.5102	0.9615	0.6667	26	0.9615	0.6410	0.7692	39	0.9167	0.5789	0.7097	19	1.0	0.7692	0.8696	13	0.7320	0.8471	0.7377	0.7538	97	0.8369	0.7320	0.7435	97	0.9374	[[0, 1, 2, 3], [0, 25, 1, 0, 0], [1, 13, 25, 1, 0], [2, 8, 0, 11, 0], [3, 3, 0, 0, 10]]
0.9214	45.82	1100	0.9620	0.9615	0.9615	0.9615	26	0.9730	0.9231	0.9474	39	0.9048	1.0	0.9500	19	1.0	1.0	1.0	13	0.9588	0.9598	0.9712	0.9647	97	0.9602	0.9588	0.9587	97	0.9328	[[0, 1, 2, 3], [0, 25, 1, 0, 0], [1, 1, 36, 2, 0], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]
0.9305	49.98	1200	0.9736	0.8125	1.0	0.8966	26	1.0	0.8205	0.9014	39	0.9048	1.0	0.9500	19	1.0	0.9231	0.9600	13	0.9175	0.9293	0.9359	0.9270	97	0.9311	0.9175	0.9175	97	0.9253	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 5, 32, 2, 0], [2, 0, 0, 19, 0], [3, 1, 0, 0, 12]]
0.8982	54.16	1300	0.9586	0.7812	0.9615	0.8621	26	0.9688	0.7949	0.8732	39	0.9	0.9474	0.9231	19	1.0	1.0	1.0	13	0.8969	0.9125	0.9259	0.9146	97	0.9092	0.8969	0.8970	97	0.9283	[[0, 1, 2, 3], [0, 25, 1, 0, 0], [1, 6, 31, 2, 0], [2, 1, 0, 18, 0], [3, 0, 0, 0, 13]]
0.8382	58.33	1400	0.8864	0.9615	0.9615	0.9615	26	0.9722	0.8974	0.9333	39	0.95	1.0	0.9744	19	0.8667	1.0	0.9286	13	0.9485	0.9376	0.9647	0.9495	97	0.9509	0.9485	0.9483	97	0.8904	[[0, 1, 2, 3], [0, 25, 1, 0, 0], [1, 1, 35, 1, 2], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]
0.7314	62.49	1500	0.7880	0.96	0.9231	0.9412	26	0.9474	0.9231	0.9351	39	0.95	1.0	0.9744	19	0.9286	1.0	0.9630	13	0.9485	0.9465	0.9615	0.9534	97	0.9488	0.9485	0.9481	97	0.8020	[[0, 1, 2, 3], [0, 24, 2, 0, 0], [1, 1, 36, 1, 1], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]
0.448	66.65	1600	0.3458	0.9615	0.9615	0.9615	26	0.9730	0.9231	0.9474	39	1.0	1.0	1.0	19	0.8667	1.0	0.9286	13	0.9588	0.9503	0.9712	0.9594	97	0.9610	0.9588	0.9590	97	0.2561	[[0, 1, 2, 3], [0, 25, 1, 0, 0], [1, 1, 36, 0, 2], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]
0.1921	70.82	1700	0.1970	0.9615	0.9615	0.9615	26	0.9730	0.9231	0.9474	39	1.0	1.0	1.0	19	0.8667	1.0	0.9286	13	0.9588	0.9503	0.9712	0.9594	97	0.9610	0.9588	0.9590	97	0.1581	[[0, 1, 2, 3], [0, 25, 1, 0, 0], [1, 1, 36, 0, 2], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]
0.1499	74.98	1800	0.1463	0.9615	0.9615	0.9615	26	0.9730	0.9231	0.9474	39	1.0	1.0	1.0	19	0.8667	1.0	0.9286	13	0.9588	0.9503	0.9712	0.9594	97	0.9610	0.9588	0.9590	97	0.1384	[[0, 1, 2, 3], [0, 25, 1, 0, 0], [1, 1, 36, 0, 2], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]
0.1099	79.16	1900	0.1459	0.9630	1.0	0.9811	26	1.0	0.9231	0.9600	39	1.0	1.0	1.0	19	0.8667	1.0	0.9286	13	0.9691	0.9574	0.9808	0.9674	97	0.9722	0.9691	0.9693	97	0.1293	[[0, 1, 2, 3], [0, 26, 0, 0, 0], [1, 1, 36, 0, 2], [2, 0, 0, 19, 0], [3, 0, 0, 0, 13]]

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu116
• Datasets 2.8.0
• Tokenizers 0.13.2