whisper-small-yue-fold-1

This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

Training Loss	Epoch	Step	Validation Loss	Wer
2.7642	0.38	50	2.3176	130.8905
0.7888	0.75	100	0.4103	233.1169
0.2973	1.13	150	0.3161	125.8813
0.2804	1.5	200	0.2891	109.2764
0.2811	1.88	250	0.2728	96.7532
0.1662	2.26	300	0.2678	108.2560
0.1532	2.63	350	0.2684	90.3525
0.1504	3.01	400	0.2573	75.7885
0.073	3.38	450	0.2696	69.1095
0.0851	3.76	500	0.2676	83.2096
0.0437	4.14	550	0.2794	67.4397
0.0428	4.51	600	0.2791	74.8609
0.0464	4.89	650	0.2767	78.6642
0.0223	5.26	700	0.2789	70.6865
0.027	5.64	750	0.2914	64.4712
0.0199	6.02	800	0.2805	80.7050
0.0117	6.39	850	0.3035	66.3265
0.0137	6.77	900	0.3083	81.3544
0.0078	7.14	950	0.3059	68.1818
0.009	7.52	1000	0.3235	70.0371
0.01	7.89	1050	0.3171	69.2950
0.0049	8.27	1100	0.3295	72.5417
0.0052	8.65	1150	0.3386	71.9852
0.0047	9.02	1200	0.3320	75.6030
0.0041	9.4	1250	0.3524	73.0056
0.0055	9.77	1300	0.3482	76.5306
0.0032	10.15	1350	0.3531	71.1503
0.0024	10.53	1400	0.3629	68.0891
0.0022	10.9	1450	0.3659	67.9963
0.0009	11.28	1500	0.3709	73.1911
0.0027	11.65	1550	0.3827	70.6865
0.0008	12.03	1600	0.3875	72.9128
0.0011	12.41	1650	0.3930	73.3766
0.0024	12.78	1700	0.3975	73.1911
0.0047	13.16	1750	0.4083	73.6549