mms_1b_audio_books_aze

hajili/mms_1b_audio_books_aze

mms_1b_audio_books_aze is a automatic speech recognition model from hajili. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as cc-by-nc-4.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

13% of all-time downloads

All-time downloads

Public

Parameters

965M

143 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors3.9 GB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: cc-by-nc-4.0
Model type: wav2vec2
Access: Public
Created: Mar 26, 2024
Updated: Mar 26, 2024
SHA: 65d708b0

Base models

facebook/mms-1b-all

Model card

From the Hugging Face model README

mms_1b_audio_books_aze

This model is a fine-tuned version of facebook/mms-1b-all on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.1030
Wer: 0.1235
Cer: 0.0240
Wer Book: 0.0991
Cer Book: 0.0165
Wer Cv: 0.2070
Cer Cv: 0.0468

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: 16
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 20
num_epochs: 10
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer	Wer Book	Cer Book	Wer Cv	Cer Cv
0.3702	0.25	159	0.1205	0.1489	0.0280	0.1327	0.0216	0.2044	0.0474
0.347	0.5	318	0.1120	0.1407	0.0263	0.1214	0.0198	0.2070	0.0461
0.3259	0.75	477	0.1181	0.1495	0.0275	0.1289	0.0208	0.2199	0.0479
0.3111	1.0	636	0.1173	0.1545	0.0284	0.1357	0.0223	0.2186	0.0472
0.3152	1.25	795	0.1085	0.1346	0.0249	0.1168	0.0180	0.1953	0.0463
0.2986	1.5	954	0.1063	0.1396	0.0259	0.1198	0.0190	0.2070	0.0470
0.2997	1.76	1113	0.1115	0.1428	0.0264	0.1206	0.0191	0.2186	0.0488
0.2867	2.01	1272	0.1107	0.1445	0.0266	0.1210	0.0195	0.2251	0.0484
0.303	2.26	1431	0.1136	0.1434	0.0268	0.1195	0.0191	0.2251	0.0502
0.2736	2.51	1590	0.1106	0.1416	0.0263	0.1183	0.0188	0.2212	0.0493
0.2777	2.76	1749	0.1088	0.1378	0.0259	0.1214	0.0194	0.1940	0.0461
0.327	3.01	1908	0.1115	0.1366	0.0259	0.1130	0.0184	0.2173	0.0486
0.2789	3.26	2067	0.1064	0.1363	0.0256	0.1176	0.0189	0.2005	0.0463
0.2828	3.51	2226	0.1066	0.1355	0.0252	0.1157	0.0189	0.2031	0.0449
0.2768	3.76	2385	0.1050	0.1302	0.0244	0.1085	0.0177	0.2044	0.0449
0.2858	4.01	2544	0.1025	0.1290	0.0240	0.1066	0.0172	0.2057	0.0449
0.2785	4.26	2703	0.1055	0.1334	0.0246	0.1096	0.0173	0.2147	0.0470
0.2733	4.51	2862	0.1055	0.1334	0.0249	0.1138	0.0177	0.2005	0.0470
0.2624	4.76	3021	0.1050	0.1290	0.0243	0.1078	0.0173	0.2018	0.0458
0.2676	5.02	3180	0.1045	0.1305	0.0241	0.1055	0.0167	0.2160	0.0470
0.2457	5.27	3339	0.1015	0.1287	0.0237	0.1074	0.0170	0.2018	0.0440
0.2637	5.52	3498	0.1030	0.1290	0.0237	0.1055	0.0168	0.2096	0.0449
0.2646	5.77	3657	0.1047	0.1346	0.0247	0.1115	0.0173	0.2135	0.0474
0.2831	6.02	3816	0.1072	0.1302	0.0245	0.1085	0.0173	0.2044	0.0465
0.2502	6.27	3975	0.1047	0.1267	0.0241	0.1043	0.0166	0.2031	0.0470
0.2427	6.52	4134	0.1043	0.1299	0.0244	0.1066	0.0168	0.2096	0.0477
0.2576	6.77	4293	0.1046	0.1249	0.0242	0.1021	0.0169	0.2031	0.0468
0.2466	7.02	4452	0.1074	0.1308	0.0246	0.1043	0.0168	0.2212	0.0484
0.2463	7.27	4611	0.1045	0.1255	0.0239	0.0983	0.0158	0.2186	0.0488
0.2377	7.52	4770	0.1047	0.1296	0.0246	0.1025	0.0167	0.2225	0.0488
0.2505	7.77	4929	0.1030	0.1279	0.0242	0.1032	0.0168	0.2122	0.0468
0.2479	8.03	5088	0.1019	0.1191	0.0229	0.0938	0.0155	0.2057	0.0458
0.2451	8.28	5247	0.1027	0.1240	0.0238	0.1006	0.0165	0.2044	0.0463
0.2442	8.53	5406	0.1026	0.1261	0.0239	0.1009	0.0164	0.2122	0.0468
0.2226	8.78	5565	0.1031	0.1246	0.0239	0.1013	0.0167	0.2044	0.0461
0.2462	9.03	5724	0.1030	0.1252	0.0239	0.1006	0.0165	0.2096	0.0467
0.2439	9.28	5883	0.1033	0.1243	0.0239	0.1009	0.0165	0.2044	0.0467
0.2363	9.53	6042	0.1036	0.1220	0.0236	0.0960	0.0159	0.2109	0.0470
0.2221	9.78	6201	0.1030	0.1235	0.0240	0.0991	0.0165	0.2070	0.0468

Framework versions

Transformers 4.40.0.dev0
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.15.2

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: cc-by-nc-4.0
Created: Mar 26, 2024
Updated: Mar 26, 2024