This model is a fine-tuned version of facebook/mms-1b-fl102 on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

Loss: 0.3051
Wer: 0.5393

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: 32
eval_batch_size: 8
seed: 42
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: 100
num_epochs: 25
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
8.748	1.5625	100	0.4114	0.6370
0.4627	3.125	200	0.3346	0.6006
0.3883	4.6875	300	0.3143	0.5725
0.3596	6.25	400	0.3133	0.5709
0.3294	7.8125	500	0.3069	0.5603
0.3078	9.375	600	0.3073	0.5516
0.2881	10.9375	700	0.3110	0.5522
0.2755	12.5	800	0.3041	0.5519
0.2627	14.0625	900	0.3163	0.5467
0.245	15.625	1000	0.3009	0.5432
0.2303	17.1875	1100	0.3074	0.5374
0.2233	18.75	1200	0.3123	0.5413
0.2142	20.3125	1300	0.3123	0.5397
0.2125	21.875	1400	0.3088	0.5403
0.2025	23.4375	1500	0.3055	0.5416
0.2072	25.0	1600	0.3051	0.5393

Framework versions

Transformers 4.46.0.dev0
Pytorch 2.4.0+cu121
Datasets 3.0.1
Tokenizers 0.20.1

abatpool/mal-mms

mal-mms

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

abatpool/
mal-mms