Helldivers2ASR_V4

This model is a fine-tuned version of facebook/wav2vec2-large-960h on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 23.9309
• Wer: 0.0418

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: 8e-05
• 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: constant
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 60
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
860.6827	1.0	505	518.6855	0.4561
646.7485	2.0	1010	404.5416	0.3735
550.9087	3.0	1515	347.0684	0.3078
479.678	4.0	2020	319.9583	0.2773
425.8765	5.0	2525	278.4693	0.2440
384.8481	6.0	3030	255.8700	0.2202
347.8814	7.0	3535	194.1332	0.1985
316.2373	8.0	4040	186.0174	0.1867
285.9741	9.0	4545	162.6776	0.1665
265.6298	10.0	5050	136.8190	0.1377
246.6871	11.0	5555	115.5616	0.1296
220.2408	12.0	6060	119.0393	0.1281
206.048	13.0	6565	105.3089	0.1142
187.9322	14.0	7070	104.7907	0.1088
179.1001	15.0	7575	84.0504	0.0909
166.2595	16.0	8080	77.4637	0.1009
160.3703	17.0	8585	70.2956	0.0825
146.5175	18.0	9090	65.9802	0.0793
135.1917	19.0	9595	69.4915	0.0824
129.9824	20.0	10100	60.3350	0.0807
128.6042	21.0	10605	56.4799	0.0743
119.3818	22.0	11110	58.4327	0.0745
111.8126	23.0	11615	49.4417	0.0733
104.4044	24.0	12120	50.4322	0.0726
105.2895	25.0	12625	48.5079	0.0738
98.1706	26.0	13130	39.8806	0.0627
96.8815	27.0	13635	43.9033	0.0683
94.4319	28.0	14140	40.7528	0.0644
88.7985	29.0	14645	38.5910	0.0841
89.5256	30.0	15150	45.4595	0.0807
86.8755	31.0	15655	43.6640	0.0709
83.8043	32.0	16160	41.5719	0.0579
80.2981	33.0	16665	32.2402	0.0632
80.228	34.0	17170	37.0875	0.0680
75.8638	35.0	17675	34.1005	0.0565
73.154	36.0	18180	33.2692	0.0637
68.464	37.0	18685	31.7566	0.0545
67.3167	38.0	19190	31.6638	0.0514
63.2984	39.0	19695	28.8354	0.0512
62.8066	40.0	20200	28.3509	0.0551
63.414	41.0	20705	24.4108	0.0504
64.3067	42.0	21210	26.4849	0.0569
57.9233	43.0	21715	24.2229	0.0587
59.5782	44.0	22220	22.2866	0.0617
56.6089	45.0	22725	26.2410	0.0444
59.7503	46.0	23230	30.5061	0.0584
54.7057	47.0	23735	27.0451	0.0512
57.7382	48.0	24240	21.9726	0.0521
50.2926	49.0	24745	26.2113	0.0534
52.8126	50.0	25250	28.5752	0.0515
50.02	51.0	25755	25.0730	0.0512
50.0685	52.0	26260	23.9120	0.0519
49.0512	53.0	26765	27.2960	0.0603
54.5931	54.0	27270	23.2472	0.0546
46.7353	55.0	27775	25.7537	0.0546
48.7158	56.0	28280	28.0895	0.0447
47.141	57.0	28785	23.7680	0.0462
46.1992	58.0	29290	21.9848	0.0413
44.0671	59.0	29795	26.6304	0.0507
44.9467	60.0	30300	23.9309	0.0418

Framework versions

• Transformers 4.44.0
• Pytorch 2.5.1+cu121
• Datasets 3.6.0
• Tokenizers 0.19.1