ft_1

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0074
• Cer: 0.0011

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
9.6728	0.08	500	0.9218	1.0
0.6369	0.17	1000	0.6685	1.0
0.511	0.25	1500	0.4867	0.5464
0.4632	0.34	2000	0.4821	0.2281
0.4354	0.42	2500	0.3887	0.1763
0.3664	0.51	3000	0.2324	0.0484
0.2641	0.59	3500	0.1729	0.0276
0.218	0.67	4000	0.1485	0.0213
0.1938	0.76	4500	0.1270	0.0188
0.1867	0.84	5000	0.1192	0.0177
0.1756	0.93	5500	0.1047	0.0163
0.1659	1.01	6000	0.1008	0.0150
0.1522	1.1	6500	0.0970	0.0140
0.1456	1.18	7000	0.0944	0.0140
0.1431	1.26	7500	0.0905	0.0134
0.1416	1.35	8000	0.0919	0.0120
0.1356	1.43	8500	0.0792	0.0124
0.1345	1.52	9000	0.0880	0.0129
0.1316	1.6	9500	0.0753	0.0111
0.1293	1.69	10000	0.0728	0.0106
0.1238	1.77	10500	0.0682	0.0109
0.121	1.85	11000	0.0645	0.0100
0.1161	1.94	11500	0.0658	0.0094
0.1107	2.02	12000	0.0669	0.0115
0.1054	2.11	12500	0.0597	0.0093
0.104	2.19	13000	0.0572	0.0088
0.1007	2.28	13500	0.0522	0.0082
0.0986	2.36	14000	0.0575	0.0090
0.103	2.44	14500	0.0513	0.0085
0.0973	2.53	15000	0.0597	0.0081
0.0949	2.61	15500	0.0474	0.0074
0.092	2.7	16000	0.0538	0.0083
0.0944	2.78	16500	0.0471	0.0078
0.0896	2.87	17000	0.0417	0.0064
0.0895	2.95	17500	0.0406	0.0063
0.0866	3.03	18000	0.0393	0.0064
0.0795	3.12	18500	0.0377	0.0067
0.0803	3.2	19000	0.0380	0.0063
0.0771	3.29	19500	0.0386	0.0058
0.0742	3.37	20000	0.0347	0.0057
0.0743	3.46	20500	0.0363	0.0062
0.073	3.54	21000	0.0314	0.0051
0.0727	3.62	21500	0.0293	0.0050
0.0693	3.71	22000	0.0331	0.0057
0.0689	3.79	22500	0.0306	0.0050
0.0712	3.88	23000	0.0304	0.0049
0.0697	3.96	23500	0.0287	0.0050
0.0656	4.05	24000	0.0269	0.0046
0.0624	4.13	24500	0.0282	0.0048
0.0614	4.21	25000	0.0262	0.0046
0.0589	4.3	25500	0.0245	0.0044
0.0596	4.38	26000	0.0237	0.0043
0.0589	4.47	26500	0.0233	0.0045
0.059	4.55	27000	0.0275	0.0044
0.0564	4.64	27500	0.0241	0.0042
0.0589	4.72	28000	0.0209	0.0044
0.055	4.8	28500	0.0199	0.0039
0.0629	4.89	29000	0.0223	0.0046
0.0581	4.97	29500	0.0234	0.0044
0.0529	5.06	30000	0.0197	0.0037
0.0489	5.14	30500	0.0232	0.0044
0.0512	5.23	31000	0.0202	0.0039
0.0501	5.31	31500	0.0182	0.0036
0.0494	5.39	32000	0.0181	0.0034
0.049	5.48	32500	0.0173	0.0034
0.0482	5.56	33000	0.0173	0.0034
0.0505	5.65	33500	0.0172	0.0033
0.048	5.73	34000	0.0168	0.0034
0.0485	5.82	34500	0.0162	0.0032
0.0461	5.9	35000	0.0154	0.0032
0.0436	5.98	35500	0.0152	0.0033
0.043	6.07	36000	0.0157	0.0033
0.0429	6.15	36500	0.0160	0.0036
0.0402	6.24	37000	0.0169	0.0034
0.043	6.32	37500	0.0135	0.0030
0.0409	6.41	38000	0.0167	0.0033
0.0446	6.49	38500	0.0155	0.0035
0.0402	6.57	39000	0.0145	0.0029
0.0427	6.66	39500	0.0139	0.0028
0.0392	6.74	40000	0.0148	0.0028
0.0395	6.83	40500	0.0144	0.0026
0.0396	6.91	41000	0.0208	0.0035
0.0394	7.0	41500	0.0127	0.0026
0.0367	7.08	42000	0.0129	0.0023
0.0368	7.16	42500	0.0120	0.0024
0.0345	7.25	43000	0.0119	0.0027
0.0342	7.33	43500	0.0123	0.0026
0.0368	7.42	44000	0.0109	0.0024
0.0358	7.5	44500	0.0120	0.0026
0.036	7.59	45000	0.0110	0.0026
0.0332	7.67	45500	0.0115	0.0024
0.0356	7.75	46000	0.0128	0.0024
0.0353	7.84	46500	0.0123	0.0024
0.0365	7.92	47000	0.0106	0.0023
0.0341	8.01	47500	0.0110	0.0023
0.0286	8.09	48000	0.0109	0.0023
0.0324	8.18	48500	0.0099	0.0022
0.0321	8.26	49000	0.0111	0.0023
0.0333	8.34	49500	0.0110	0.0023
0.0302	8.43	50000	0.0116	0.0026
0.0325	8.51	50500	0.0123	0.0025
0.0317	8.6	51000	0.0101	0.0027
0.032	8.68	51500	0.0108	0.0024
0.0298	8.77	52000	0.0105	0.0024
0.0304	8.85	52500	0.0113	0.0022
0.0315	8.93	53000	0.0109	0.0023
0.0299	9.02	53500	0.0102	0.0022
0.029	9.1	54000	0.0108	0.0021
0.0261	9.19	54500	0.0112	0.0022
0.0286	9.27	55000	0.0116	0.0023
0.0293	9.36	55500	0.0108	0.0023
0.0267	9.44	56000	0.0109	0.0023
0.0271	9.52	56500	0.0103	0.0021
0.0276	9.61	57000	0.0102	0.0021
0.0256	9.69	57500	0.0101	0.0021
0.0264	9.78	58000	0.0131	0.0021
0.0277	9.86	58500	0.0096	0.0021
0.0303	9.95	59000	0.0097	0.0020
0.027	10.03	59500	0.0102	0.0021
0.0241	10.11	60000	0.0089	0.0020
0.0226	10.2	60500	0.0108	0.0020
0.0248	10.28	61000	0.0109	0.0021
0.0244	10.37	61500	0.0093	0.0020
0.025	10.45	62000	0.0167	0.0021
0.0238	10.54	62500	0.0123	0.0019
0.0253	10.62	63000	0.0089	0.0019
0.0247	10.7	63500	0.0090	0.0019
0.0229	10.79	64000	0.0084	0.0017
0.0242	10.87	64500	0.0086	0.0019
0.0251	10.96	65000	0.0082	0.0017
0.0222	11.04	65500	0.0081	0.0017
0.0233	11.13	66000	0.0078	0.0018
0.0226	11.21	66500	0.0081	0.0017
0.0227	11.29	67000	0.0096	0.0018
0.024	11.38	67500	0.0081	0.0017
0.0217	11.46	68000	0.0088	0.0019
0.0222	11.55	68500	0.0086	0.0019
0.0229	11.63	69000	0.0091	0.0021
0.02	11.72	69500	0.0094	0.0020
0.022	11.8	70000	0.0095	0.0017
0.0232	11.88	70500	0.0078	0.0018
0.0216	11.97	71000	0.0088	0.0017
0.0216	12.05	71500	0.0081	0.0017
0.0182	12.14	72000	0.0088	0.0018
0.0196	12.22	72500	0.0092	0.0017
0.0203	12.31	73000	0.0083	0.0018
0.0191	12.39	73500	0.0086	0.0018
0.0192	12.47	74000	0.0122	0.0016
0.0194	12.56	74500	0.0084	0.0016
0.0192	12.64	75000	0.0098	0.0016
0.0192	12.73	75500	0.0101	0.0017
0.0191	12.81	76000	0.0100	0.0016
0.0191	12.9	76500	0.0091	0.0016
0.0184	12.98	77000	0.0084	0.0017
0.0188	13.06	77500	0.0080	0.0016
0.0188	13.15	78000	0.0094	0.0017
0.0172	13.23	78500	0.0098	0.0017
0.0174	13.32	79000	0.0134	0.0015
0.0177	13.4	79500	0.0106	0.0016
0.0178	13.49	80000	0.0100	0.0014
0.017	13.57	80500	0.0104	0.0016
0.0169	13.65	81000	0.0094	0.0014
0.0189	13.74	81500	0.0089	0.0015
0.0172	13.82	82000	0.0086	0.0016
0.0167	13.91	82500	0.0091	0.0015
0.0179	13.99	83000	0.0088	0.0015
0.0175	14.08	83500	0.0076	0.0014
0.0164	14.16	84000	0.0082	0.0013
0.0143	14.24	84500	0.0080	0.0015
0.0158	14.33	85000	0.0082	0.0014
0.0153	14.41	85500	0.0086	0.0016
0.0173	14.5	86000	0.0077	0.0015
0.0149	14.58	86500	0.0084	0.0016
0.0153	14.67	87000	0.0078	0.0015
0.0157	14.75	87500	0.0074	0.0014
0.0173	14.83	88000	0.0086	0.0015
0.0171	14.92	88500	0.0080	0.0014
0.0148	15.0	89000	0.0073	0.0013
0.0145	15.09	89500	0.0074	0.0014
0.0153	15.17	90000	0.0068	0.0014
0.0156	15.26	90500	0.0070	0.0014
0.0143	15.34	91000	0.0072	0.0014
0.0143	15.42	91500	0.0069	0.0015
0.0139	15.51	92000	0.0070	0.0014
0.0149	15.59	92500	0.0087	0.0015
0.0124	15.68	93000	0.0074	0.0013
0.0139	15.76	93500	0.0076	0.0013
0.0148	15.85	94000	0.0074	0.0014
0.013	15.93	94500	0.0073	0.0014
0.0138	16.01	95000	0.0069	0.0013
0.0148	16.1	95500	0.0069	0.0013
0.0135	16.18	96000	0.0067	0.0014
0.0143	16.27	96500	0.0067	0.0012
0.014	16.35	97000	0.0071	0.0013
0.0138	16.44	97500	0.0072	0.0012
0.0132	16.52	98000	0.0070	0.0012
0.0122	16.6	98500	0.0077	0.0013
0.0124	16.69	99000	0.0078	0.0013
0.0138	16.77	99500	0.0070	0.0012
0.0118	16.86	100000	0.0071	0.0013
0.0123	16.94	100500	0.0068	0.0013
0.0113	17.03	101000	0.0070	0.0013
0.011	17.11	101500	0.0076	0.0013
0.0117	17.19	102000	0.0073	0.0013
0.0123	17.28	102500	0.0078	0.0012
0.0132	17.36	103000	0.0072	0.0013
0.0105	17.45	103500	0.0078	0.0012
0.0115	17.53	104000	0.0078	0.0012
0.0112	17.62	104500	0.0076	0.0012
0.0108	17.7	105000	0.0080	0.0012
0.0106	17.78	105500	0.0080	0.0012
0.0114	17.87	106000	0.0074	0.0012
0.0127	17.95	106500	0.0070	0.0012
0.0105	18.04	107000	0.0073	0.0012
0.0108	18.12	107500	0.0083	0.0012
0.0107	18.21	108000	0.0088	0.0012
0.0102	18.29	108500	0.0080	0.0011
0.0089	18.37	109000	0.0073	0.0011
0.0107	18.46	109500	0.0073	0.0011
0.0112	18.54	110000	0.0078	0.0011
0.0106	18.63	110500	0.0077	0.0011
0.0101	18.71	111000	0.0078	0.0011
0.0111	18.8	111500	0.0080	0.0011
0.0106	18.88	112000	0.0073	0.0011
0.0102	18.96	112500	0.0075	0.0011
0.0098	19.05	113000	0.0075	0.0011
0.0118	19.13	113500	0.0078	0.0011
0.0094	19.22	114000	0.0077	0.0011
0.009	19.3	114500	0.0077	0.0011
0.0094	19.39	115000	0.0077	0.0011
0.0084	19.47	115500	0.0074	0.0010
0.011	19.55	116000	0.0076	0.0011
0.0106	19.64	116500	0.0073	0.0011
0.009	19.72	117000	0.0074	0.0011
0.0097	19.81	117500	0.0074	0.0011
0.0095	19.89	118000	0.0074	0.0011
0.0094	19.98	118500	0.0074	0.0011

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.1+cu121
• Datasets 2.13.0
• Tokenizers 0.15.0