my_awesome_asr_mind_model_2

This model is a fine-tuned version of facebook/wav2vec2-base on the minds14 dataset. It achieves the following results on the evaluation set:

• Loss: 11.9803
• Wer: 1.0

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.3509	10	72.3784	1.0325
No log	0.7018	20	69.9115	1.0282
120.859	1.0351	30	62.1146	1.0276
120.859	1.3860	40	68.2888	1.0393
123.4983	1.7368	50	62.4194	1.0319
123.4983	2.0702	60	63.7456	1.0282
123.4983	2.4211	70	61.1313	1.0362
114.5327	2.7719	80	57.3109	1.0393
114.5327	3.1053	90	63.8890	1.0405
92.0841	3.4561	100	54.7270	1.0368
92.0841	3.8070	110	52.6792	1.0166
92.0841	4.1404	120	48.9400	1.0043
82.4037	4.4912	130	45.7991	1.0006
82.4037	4.8421	140	42.2452	1.0
66.3204	5.1754	150	42.3656	1.0
66.3204	5.5263	160	35.6008	1.0
66.3204	5.8772	170	33.9067	1.0
54.6729	6.2105	180	35.5143	1.0
54.6729	6.5614	190	29.4490	1.0
47.0622	6.9123	200	31.8126	1.0
47.0622	7.2456	210	29.0946	1.0
47.0622	7.5965	220	29.9088	1.0
45.3162	7.9474	230	31.5442	1.0
45.3162	8.2807	240	32.5329	1.0
42.1556	8.6316	250	23.7687	1.0
42.1556	8.9825	260	26.9307	1.0
42.1556	9.3158	270	28.8010	1.0
37.6117	9.6667	280	25.4160	1.0
37.6117	10.0	290	31.1950	1.0
44.1981	10.3509	300	27.4846	1.0
44.1981	10.7018	310	27.9379	1.0
44.1981	11.0351	320	29.6825	1.0
37.2733	11.3860	330	28.1058	1.0
37.2733	11.7368	340	32.6282	1.0
43.2232	12.0702	350	26.3374	1.0
43.2232	12.4211	360	27.2462	1.0
43.2232	12.7719	370	29.3861	1.0
39.612	13.1053	380	35.5876	1.0
39.612	13.4561	390	26.4560	1.0
38.569	13.8070	400	30.3207	1.0
38.569	14.1404	410	24.5547	1.0
38.569	14.4912	420	17.0083	1.0
32.3299	14.8421	430	14.2530	1.0
32.3299	15.1754	440	12.6302	1.0
20.7078	15.5263	450	13.9304	1.0
20.7078	15.8772	460	11.8625	1.0
20.7078	16.2105	470	9.0681	1.0
18.1266	16.5614	480	11.8818	1.0
18.1266	16.9123	490	11.3368	1.0
18.9089	17.2456	500	10.6305	1.0
18.9089	17.5965	510	12.9240	1.0
18.9089	17.9474	520	11.0008	1.0
16.9312	18.2807	530	11.0118	1.0
16.9312	18.6316	540	12.8748	1.0
16.9932	18.9825	550	11.6355	1.0
16.9932	19.3158	560	10.7317	1.0
16.9932	19.6667	570	13.6561	1.0
16.7623	20.0	580	11.6266	1.0
16.7623	20.3509	590	11.2916	1.0
17.8217	20.7018	600	11.7152	1.0
17.8217	21.0351	610	9.4275	1.0
17.8217	21.3860	620	13.7821	1.0
17.5088	21.7368	630	10.2990	1.0
17.5088	22.0702	640	9.6245	1.0
19.6078	22.4211	650	12.9627	1.0
19.6078	22.7719	660	15.1354	1.0
19.6078	23.1053	670	9.5284	1.0
16.1741	23.4561	680	12.0476	1.0
16.1741	23.8070	690	11.7502	1.0
17.9308	24.1404	700	10.6731	1.0
17.9308	24.4912	710	7.9770	1.0
17.9308	24.8421	720	12.0266	1.0
14.9285	25.1754	730	12.6227	1.0
14.9285	25.5263	740	13.2743	1.0
16.4239	25.8772	750	15.2944	1.0
16.4239	26.2105	760	10.8676	1.0
16.4239	26.5614	770	10.1779	1.0
16.4257	26.9123	780	9.8973	1.0
16.4257	27.2456	790	9.1979	1.0
15.2576	27.5965	800	10.5384	1.0
15.2576	27.9474	810	10.2605	1.0
15.2576	28.2807	820	10.1279	1.0
16.9401	28.6316	830	14.6968	1.0
16.9401	28.9825	840	11.0547	1.0
16.8678	29.3158	850	9.5137	1.0
16.8678	29.6667	860	8.0353	1.0
16.8678	30.0	870	9.4884	1.0
14.1705	30.3509	880	7.4831	1.0
14.1705	30.7018	890	10.3853	1.0
12.4927	31.0351	900	9.0745	1.0
12.4927	31.3860	910	12.7624	1.0
12.4927	31.7368	920	10.9480	1.0
13.2511	32.0702	930	9.6001	1.0
13.2511	32.4211	940	10.0557	1.0
11.6225	32.7719	950	9.7758	1.0
11.6225	33.1053	960	9.1229	1.0
11.6225	33.4561	970	10.5019	1.0
9.8833	33.8070	980	10.4892	1.0
9.8833	34.1404	990	11.0348	1.0
9.4119	34.4912	1000	8.3344	1.0
9.4119	34.8421	1010	7.9083	1.0
9.4119	35.1754	1020	12.6228	1.0
9.2786	35.5263	1030	10.6666	1.0
9.2786	35.8772	1040	8.6810	1.0
8.7912	36.2105	1050	6.8696	1.0006
8.7912	36.5614	1060	7.5894	1.0
8.7912	36.9123	1070	10.2622	1.0
8.7841	37.2456	1080	5.6630	1.0
8.7841	37.5965	1090	11.6318	1.0
8.6711	37.9474	1100	8.7652	1.0018
8.6711	38.2807	1110	8.7590	1.0
8.6711	38.6316	1120	12.6484	1.0
8.539	38.9825	1130	9.6365	1.0
8.539	39.3158	1140	9.4819	1.0
8.1748	39.6667	1150	12.8334	1.0
8.1748	40.0	1160	13.2719	1.0
8.1748	40.3509	1170	8.2957	1.0
8.1309	40.7018	1180	12.5595	1.0
8.1309	41.0351	1190	11.5797	1.0
8.1256	41.3860	1200	11.6983	1.0
8.1256	41.7368	1210	9.5951	1.0
8.1256	42.0702	1220	11.3728	1.0
7.8759	42.4211	1230	8.7616	1.0
7.8759	42.7719	1240	9.5483	1.0
7.7428	43.1053	1250	9.1343	1.0
7.7428	43.4561	1260	8.3384	1.0
7.7428	43.8070	1270	7.6469	1.0
7.8274	44.1404	1280	13.3626	1.0
7.8274	44.4912	1290	10.8081	1.0
7.7745	44.8421	1300	7.2929	1.0
7.7745	45.1754	1310	13.2838	1.0
7.7745	45.5263	1320	9.0075	1.0
7.5423	45.8772	1330	8.6317	1.0
7.5423	46.2105	1340	6.8053	1.0049
7.8556	46.5614	1350	10.9606	1.0
7.8556	46.9123	1360	11.3875	1.0
7.8556	47.2456	1370	8.8268	1.0
7.435	47.5965	1380	8.9938	1.0
7.435	47.9474	1390	7.1973	1.0
7.5196	48.2807	1400	10.2050	1.0
7.5196	48.6316	1410	7.6926	1.0
7.5196	48.9825	1420	10.3563	1.0
7.2988	49.3158	1430	10.2332	1.0
7.2988	49.6667	1440	9.7126	1.0
7.9996	50.0	1450	10.2664	1.0
7.9996	50.3509	1460	11.0007	1.0
7.9996	50.7018	1470	7.6255	1.0
7.4441	51.0351	1480	12.5091	1.0
7.4441	51.3860	1490	8.5649	1.0
7.158	51.7368	1500	11.0922	1.0
7.158	52.0702	1510	5.5673	1.0
7.158	52.4211	1520	10.2015	1.0
7.3232	52.7719	1530	9.6431	1.0
7.3232	53.1053	1540	12.6101	1.0
7.2851	53.4561	1550	7.4011	1.0
7.2851	53.8070	1560	12.8892	1.0
7.2851	54.1404	1570	12.7549	1.0
7.13	54.4912	1580	7.0643	1.0
7.13	54.8421	1590	13.2393	1.0
7.241	55.1754	1600	7.7935	1.0
7.241	55.5263	1610	8.4647	1.0
7.241	55.8772	1620	9.1026	1.0
7.0737	56.2105	1630	7.7952	1.0
7.0737	56.5614	1640	7.5957	1.0
7.3016	56.9123	1650	11.8872	1.0
7.3016	57.2456	1660	11.6042	1.0
7.3016	57.5965	1670	8.3765	1.0
7.0452	57.9474	1680	11.1004	1.0
7.0452	58.2807	1690	15.3519	1.0
7.1912	58.6316	1700	9.1837	1.0
7.1912	58.9825	1710	9.0901	1.0
7.1912	59.3158	1720	10.7072	1.0
7.1017	59.6667	1730	11.8671	1.0
7.1017	60.0	1740	8.7281	1.0
7.0886	60.3509	1750	11.4665	1.0
7.0886	60.7018	1760	9.1317	1.0
7.0886	61.0351	1770	9.4837	1.0
7.4933	61.3860	1780	11.0652	1.0
7.4933	61.7368	1790	8.6223	1.0
6.9543	62.0702	1800	7.3605	1.0
6.9543	62.4211	1810	10.4706	1.0
6.9543	62.7719	1820	13.1088	1.0
7.1395	63.1053	1830	9.2800	1.0
7.1395	63.4561	1840	9.6354	1.0
7.0871	63.8070	1850	7.5116	1.0
7.0871	64.1404	1860	8.7661	1.0
7.0871	64.4912	1870	10.5800	1.0
6.9937	64.8421	1880	11.2561	1.0
6.9937	65.1754	1890	8.7216	1.0
6.9257	65.5263	1900	9.6056	1.0
6.9257	65.8772	1910	11.6064	1.0
6.9257	66.2105	1920	15.2734	1.0
6.9951	66.5614	1930	12.0877	1.0
6.9951	66.9123	1940	7.6033	1.0
6.8757	67.2456	1950	11.5104	1.0
6.8757	67.5965	1960	9.6040	1.0
6.8757	67.9474	1970	7.8608	1.0
6.9293	68.2807	1980	13.0502	1.0
6.9293	68.6316	1990	7.9725	1.0
7.0953	68.9825	2000	11.9803	1.0

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.1.0a0+32f93b1
• Datasets 3.2.0
• Tokenizers 0.21.0