This model is a fine-tuned version of openai/whisper-large-v2 on the mozilla-foundation/common_voice_17_0 eu dataset. It achieves the following results on the evaluation set:

Loss: 0.2702
Wer: 6.9900

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: 3.75e-05
train_batch_size: 32
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 64
optimizer: Use OptimizerNames.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: 40000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0714	2.3474	1000	0.1704	10.3476
0.0292	4.6948	2000	0.1746	9.3417
0.0141	7.0423	3000	0.1892	8.8662
0.0145	9.3897	4000	0.1899	8.9551
0.0102	11.7371	5000	0.1955	8.4127
0.0067	14.0845	6000	0.2070	9.0650
0.0061	16.4319	7000	0.2164	8.7700
0.0053	18.7793	8000	0.2156	8.4613
0.0056	21.1268	9000	0.2169	8.4952
0.0032	23.4742	10000	0.2255	8.4091
0.0072	25.8216	11000	0.2302	9.5506
0.0038	28.1690	12000	0.2232	8.7728
0.0033	30.5164	13000	0.2152	7.9538
0.0052	32.8638	14000	0.2249	8.7398
0.0014	35.2113	15000	0.2307	8.0481
0.0025	37.5587	16000	0.2272	8.2075
0.004	39.9061	17000	0.2400	8.7242
0.0011	42.2535	18000	0.2280	7.9913
0.0024	44.6009	19000	0.2388	8.8351
0.0027	46.9484	20000	0.2448	8.6821
0.0006	49.2958	21000	0.2380	7.9611
0.0005	51.6432	22000	0.2411	7.7614
0.0011	53.9906	23000	0.2360	7.8420
0.0005	56.3380	24000	0.2373	7.6927
0.0007	58.6854	25000	0.2436	8.0646
0.0007	61.0329	26000	0.2475	7.9593
0.0012	63.3803	27000	0.2484	8.3165
0.0006	65.7277	28000	0.2541	7.8805
0.0	68.0751	29000	0.2481	7.4655
0.0	70.4225	30000	0.2580	7.2044
0.0	72.7700	31000	0.2641	7.0660
0.0	75.1174	32000	0.2702	6.9900
0.0	77.4648	33000	0.2758	6.9937
0.0008	79.8122	34000	0.2651	8.2716
0.0	82.1596	35000	0.2557	7.3894
0.0008	84.5070	36000	0.2608	7.3134
0.0001	86.8545	37000	0.2644	7.1549
0.0	89.2019	38000	0.2683	7.0899
0.0	91.5493	39000	0.2711	7.0450
0.0	93.8967	40000	0.2724	7.0312

Framework versions

Transformers 4.52.3
Pytorch 2.6.0+cu124
Datasets 3.6.0
Tokenizers 0.21.1

zuazo/whisper-large-v2-eu-cv17_0

Whisper Large-V2 Basque

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

zuazo/
whisper-large-v2-eu-cv17_0