CoolFace
Modelpublic

asr-africa/bambara_mms_20_hour_jeli_asr_dataset

sourceHugging Facecc-by-nc-4.0updated 2y agoView on Hugging Face
2likes17downloads
Model Card

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>

bambaramms20hourjeliasrdataset

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

  • —Loss: 2.5960
  • —Wer: 0.1951
  • —Cer: 0.0927

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: 0.0003
  • —trainbatchsize: 8
  • —evalbatchsize: 16
  • —seed: 42
  • —gradientaccumulationsteps: 2
  • —totaltrainbatch_size: 16
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 500
  • —num_epochs: 50

Training results

Training LossEpochStepValidation LossWerCer
2.37810.43655001.77900.90070.4152
1.52710.873010001.48090.74240.3462
1.42161.309515001.41340.83160.4020
1.39951.746020001.36690.74220.3379
1.35212.182525001.25310.69690.3176
1.31612.618930001.33630.66430.3061
1.27593.055435001.29530.66070.3018
1.26043.491940001.26300.62800.2889
1.24243.928445001.46420.63080.2847
1.16394.364950001.35930.62160.2797
1.21224.801455001.20430.59660.2716
1.15045.237960001.13640.58920.2688
1.1565.674465001.14050.60490.2726
1.13716.110970001.18540.58170.2633
1.09816.547475001.13060.58760.2625
1.10216.983880001.10260.59100.2659
1.02957.420385001.25060.56710.2600
1.07027.856890001.16720.55210.2529
1.02778.293395001.09660.56160.2544
1.00238.7298100001.13890.53530.2403
0.99459.1663105001.34340.53020.2420
0.95339.6028110001.15460.53910.2517
0.967510.0393115001.19660.53550.2451
0.906110.4758120001.18080.51160.2310
0.924310.9123125001.11890.50950.2290
0.883411.3488130001.21890.49790.2226
0.881911.7852135001.20350.49100.2158
0.852212.2217140001.13850.49610.2173
0.841712.6582145001.10600.47870.2110
0.835213.0947150001.12950.49570.2237
0.785713.5312155001.09460.48140.2142
0.796313.9677160001.08910.48440.2240
0.76214.4042165001.06060.48320.2177
0.759414.8407170001.04150.45290.1992
0.736815.2772175001.08820.43990.1930
0.715815.7137180001.08720.45210.1972
0.710216.1502185001.09490.42590.1842
0.678916.5866190001.12070.41380.1821
0.689817.0231195001.12870.41050.1792
0.646317.4596200001.21310.41030.1793
0.652517.8961205001.19860.40010.1733
0.611618.3326210001.22550.40580.1778
0.613818.7691215001.20270.39460.1762
0.603319.2056220001.16810.38700.1686
0.581619.6421225001.14640.38220.1662
0.582620.0786230001.17670.38170.1651
0.550420.5151235001.28050.37960.1664
0.565620.9515240001.18950.36340.1581
0.520421.3880245001.21110.35690.1531
0.518621.8245250001.28400.35260.1541
0.507422.2610255001.21230.35640.1558
0.496622.6975260001.17400.34670.1511
0.488623.1340265001.32080.33510.1459
0.462823.5705270001.39050.32770.1439
0.474324.0070275001.33960.33780.1469
0.440824.4435280001.37670.31640.1384
0.445724.8800285001.26070.32310.1364
0.424225.3165290001.25620.31810.1383
0.427925.7529295001.25230.31980.1379
0.411626.1894300001.36250.30860.1332
0.396326.6259305001.21430.31320.1346
0.394527.0624310001.29730.29930.1320
0.373327.4989315001.35420.29550.1296
0.381627.9354320001.38040.29460.1307
0.348728.3719325001.42060.28410.1233
0.352128.8084330001.42940.28190.1236
0.335129.2449335001.56580.27970.1218
0.328529.6814340001.51030.28030.1235
0.325330.1179345001.49570.27040.1209
0.30830.5543350001.69640.26480.1173
0.318430.9908355001.47960.26090.1153
0.294131.4273360001.55270.25970.1169
0.289731.8638365001.59070.25740.1150
0.288332.3003370001.57180.25360.1132
0.279232.7368375001.55050.25270.1134
0.277333.1733380001.66070.24800.1102
0.257933.6098385001.89620.24610.1108
0.265834.0463390001.91360.24260.1116
0.253934.4828395001.91310.24400.1113
0.250134.9192400001.72900.23680.1083
0.235835.3557405001.95860.23090.1059
0.239535.7922410001.76170.22950.1040
0.226736.2287415001.87790.22640.1021
0.223936.6652420001.86960.22480.1019
0.218337.1017425001.84450.22290.1026
0.214937.5382430001.98650.22650.1041
0.212537.9747435001.89980.22520.1039
0.200138.4112440002.05910.22140.1010
0.20438.8477445001.94360.21450.1004
0.196739.2842450002.05360.21480.0994
0.190239.7206455002.05840.21610.1000
0.186240.1571460002.07440.21460.1004
0.18240.5936465002.07310.21080.0982
0.187641.0301470002.08950.20960.0978
0.174441.4666475002.23550.20380.0956
0.17841.9031480002.10990.20990.0969
0.167742.3396485002.22600.20430.0959
0.168842.7761490002.21020.20460.0949
0.167643.2126495002.21650.20470.0965
0.158943.6491500002.27410.20160.0937
0.161344.0856505002.20690.20210.0947
0.152944.5220510002.30350.20180.0958
0.151344.9585515002.47500.20070.0957
0.151545.3950520002.50790.20110.0956
0.146345.8315525002.52470.19940.0944
0.143346.2680530002.55640.19680.0934
0.14346.7045535002.52300.19700.0933
0.139947.1410540002.55320.19540.0927
0.137747.5775545002.48110.19850.0933
0.135848.0140550002.60650.19800.0934
0.137348.4505555002.58080.19510.0926
0.132748.8869560002.54580.19540.0930
0.132949.3234565002.59480.19420.0927
0.136149.7599570002.59600.19510.0927

Framework versions

  • —Transformers 4.45.1
  • —Pytorch 2.1.0+cu118
  • —Datasets 2.17.0
  • —Tokenizers 0.20.3