wav2vec2-large-mms-1b-zh-CN

This model is a fine-tuned version of facebook/mms-1b-all on the common_voice_6_1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.9552
• Cer: 0.2071

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: 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: 100
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
42.0738	0.04	100	2.9914	0.4865
2.534	0.09	200	2.0714	0.3981
2.0311	0.13	300	1.9086	0.3844
1.9237	0.17	400	1.7770	0.3650
1.865	0.22	500	1.6745	0.3579
1.8275	0.26	600	1.6277	0.3414
1.8094	0.3	700	1.6812	0.3639
1.7503	0.35	800	1.6279	0.3427
1.7448	0.39	900	1.5611	0.3376
1.7459	0.43	1000	1.5413	0.3323
1.7191	0.47	1100	1.5259	0.3280
1.6317	0.52	1200	1.5102	0.3242
1.6881	0.56	1300	1.4851	0.3212
1.6401	0.6	1400	1.4589	0.3097
1.5909	0.65	1500	1.4985	0.3186
1.618	0.69	1600	1.4415	0.3122
1.6842	0.73	1700	1.4596	0.3161
1.5413	0.78	1800	1.4275	0.3003
1.6461	0.82	1900	1.4214	0.3073
1.5536	0.86	2000	1.3924	0.3003
1.545	0.91	2100	1.3727	0.2907
1.6354	0.95	2200	1.4157	0.3088
1.4913	0.99	2300	1.4012	0.3042
1.2739	1.04	2400	1.3079	0.2855
1.2292	1.08	2500	1.3085	0.2832
1.2424	1.12	2600	1.3273	0.2879
1.2181	1.16	2700	1.3241	0.2864
1.2101	1.21	2800	1.2526	0.2780
1.26	1.25	2900	1.2949	0.2815
1.2154	1.29	3000	1.2932	0.2787
1.2446	1.34	3100	1.2774	0.2792
1.1975	1.38	3200	1.2641	0.2751
1.2048	1.42	3300	1.2645	0.2773
1.1858	1.47	3400	1.2616	0.2741
1.202	1.51	3500	1.2572	0.2725
1.1802	1.55	3600	1.2554	0.2723
1.1912	1.6	3700	1.2703	0.2657
1.213	1.64	3800	1.2491	0.2743
1.1949	1.68	3900	1.2497	0.2734
1.1813	1.73	4000	1.2367	0.2709
1.1935	1.77	4100	1.2174	0.2677
1.1842	1.81	4200	1.2307	0.2660
1.215	1.86	4300	1.2275	0.2696
1.2102	1.9	4400	1.1964	0.2595
1.2206	1.94	4500	1.2046	0.2574
1.2292	1.98	4600	1.1900	0.2595
1.034	2.03	4700	1.1849	0.2547
0.8787	2.07	4800	1.1889	0.2558
0.9124	2.11	4900	1.1809	0.2590
0.9027	2.16	5000	1.1927	0.2608
0.9158	2.2	5100	1.1860	0.2556
0.8683	2.24	5200	1.1660	0.2522
0.8932	2.29	5300	1.1477	0.2533
0.9332	2.33	5400	1.1702	0.2543
0.9427	2.37	5500	1.1653	0.2523
0.9085	2.42	5600	1.1739	0.2539
0.9238	2.46	5700	1.2005	0.2589
0.9319	2.5	5800	1.1877	0.2567
0.9414	2.55	5900	1.1730	0.2505
0.9428	2.59	6000	1.1721	0.2576
0.942	2.63	6100	1.1793	0.2547
0.9273	2.67	6200	1.1787	0.2570
0.9963	2.72	6300	1.1570	0.2540
0.9519	2.76	6400	1.1738	0.2563
0.962	2.8	6500	1.1929	0.2628
0.9765	2.85	6600	1.1531	0.2527
0.9226	2.89	6700	1.1577	0.2553
0.9492	2.93	6800	1.1490	0.2506
0.9186	2.98	6900	1.1402	0.2500
0.8681	3.02	7000	1.1520	0.2516
0.7738	3.06	7100	1.1404	0.2527
0.7605	3.11	7200	1.1535	0.2514
0.7254	3.15	7300	1.1679	0.2490
0.7422	3.19	7400	1.1536	0.2502
0.823	3.24	7500	1.1516	0.2477
0.7909	3.28	7600	1.1442	0.2459
0.7748	3.32	7700	1.1522	0.2493
0.7957	3.36	7800	1.1383	0.2470
0.7383	3.41	7900	1.1343	0.2452
0.8093	3.45	8000	1.1426	0.2467
0.8141	3.49	8100	1.1357	0.2466
0.7891	3.54	8200	1.1552	0.2480
0.8246	3.58	8300	1.1555	0.2475
0.7958	3.62	8400	1.1615	0.2502
0.7721	3.67	8500	1.1041	0.2396
0.7773	3.71	8600	1.1215	0.2411
0.7847	3.75	8700	1.1130	0.2419
0.7971	3.8	8800	1.1056	0.2469
0.7801	3.84	8900	1.1129	0.2435
0.7843	3.88	9000	1.1027	0.2387
0.7842	3.93	9100	1.0981	0.2401
0.7661	3.97	9200	1.1060	0.2428
0.7622	4.01	9300	1.0790	0.2338
0.6405	4.06	9400	1.0871	0.2352
0.6102	4.1	9500	1.0860	0.2344
0.6419	4.14	9600	1.0782	0.2356
0.6058	4.18	9700	1.0739	0.2291
0.6632	4.23	9800	1.1008	0.2366
0.6373	4.27	9900	1.0847	0.2354
0.6358	4.31	10000	1.0722	0.2313
0.6531	4.36	10100	1.0796	0.2326
0.6383	4.4	10200	1.0736	0.2322
0.6537	4.44	10300	1.0723	0.2305
0.6321	4.49	10400	1.0703	0.2329
0.6683	4.53	10500	1.0769	0.2332
0.6272	4.57	10600	1.0555	0.2292
0.651	4.62	10700	1.0570	0.2323
0.6392	4.66	10800	1.0738	0.2313
0.665	4.7	10900	1.0536	0.2276
0.677	4.75	11000	1.0554	0.2277
0.6419	4.79	11100	1.0487	0.2258
0.6549	4.83	11200	1.0427	0.2287
0.6373	4.87	11300	1.0502	0.2291
0.6642	4.92	11400	1.0411	0.2255
0.6674	4.96	11500	1.0345	0.2248
0.6733	5.0	11600	1.0440	0.2278
0.5281	5.05	11700	1.0477	0.2253
0.5465	5.09	11800	1.0553	0.2284
0.5375	5.13	11900	1.0550	0.2309
0.5103	5.18	12000	1.0433	0.2237
0.5196	5.22	12100	1.0534	0.2301
0.5645	5.26	12200	1.0492	0.2278
0.5421	5.31	12300	1.0515	0.2281
0.5234	5.35	12400	1.0383	0.2229
0.571	5.39	12500	1.0569	0.2278
0.5392	5.44	12600	1.0469	0.2253
0.5867	5.48	12700	1.0373	0.2264
0.5819	5.52	12800	1.0164	0.2237
0.5504	5.57	12900	1.0183	0.2217
0.5532	5.61	13000	1.0167	0.2232
0.5575	5.65	13100	1.0292	0.2244
0.5593	5.69	13200	1.0368	0.2247
0.5498	5.74	13300	1.0215	0.2231
0.5462	5.78	13400	1.0330	0.2212
0.5751	5.82	13500	1.0179	0.2223
0.5492	5.87	13600	1.0224	0.2202
0.5746	5.91	13700	1.0151	0.2219
0.5288	5.95	13800	1.0154	0.2199
0.5614	6.0	13900	1.0158	0.2210
0.4563	6.04	14000	1.0120	0.2197
0.502	6.08	14100	1.0125	0.2201
0.4896	6.13	14200	1.0011	0.2160
0.4774	6.17	14300	1.0027	0.2180
0.4734	6.21	14400	1.0026	0.2170
0.486	6.26	14500	0.9994	0.2177
0.4815	6.3	14600	0.9977	0.2174
0.4972	6.34	14700	1.0004	0.2175
0.4832	6.38	14800	0.9922	0.2130
0.4682	6.43	14900	0.9998	0.2167
0.4654	6.47	15000	0.9886	0.2150
0.4665	6.51	15100	0.9844	0.2154
0.4696	6.56	15200	0.9801	0.2136
0.4732	6.6	15300	0.9830	0.2145
0.4391	6.64	15400	0.9886	0.2165
0.5035	6.69	15500	0.9872	0.2157
0.4721	6.73	15600	0.9895	0.2132
0.466	6.77	15700	0.9910	0.2147
0.4981	6.82	15800	0.9934	0.2157
0.4856	6.86	15900	0.9888	0.2126
0.4798	6.9	16000	0.9830	0.2150
0.4771	6.95	16100	0.9845	0.2153
0.473	6.99	16200	0.9814	0.2116
0.4256	7.03	16300	0.9771	0.2131
0.4133	7.08	16400	0.9803	0.2125
0.4051	7.12	16500	0.9778	0.2116
0.4274	7.16	16600	0.9809	0.2116
0.4307	7.2	16700	0.9720	0.2109
0.4223	7.25	16800	0.9730	0.2109
0.4246	7.29	16900	0.9710	0.2100
0.4478	7.33	17000	0.9670	0.2101
0.4016	7.38	17100	0.9664	0.2096
0.4289	7.42	17200	0.9667	0.2093
0.4107	7.46	17300	0.9661	0.2096
0.4643	7.51	17400	0.9665	0.2106
0.433	7.55	17500	0.9673	0.2097
0.4239	7.59	17600	0.9639	0.2096
0.4144	7.64	17700	0.9635	0.2091
0.428	7.68	17800	0.9604	0.2094
0.4312	7.72	17900	0.9585	0.2099
0.4164	7.77	18000	0.9599	0.2093
0.4308	7.81	18100	0.9587	0.2080
0.4177	7.85	18200	0.9575	0.2084
0.4509	7.89	18300	0.9567	0.2082
0.4244	7.94	18400	0.9558	0.2072
0.4246	7.98	18500	0.9552	0.2071

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.14.4
• Tokenizers 0.13.3