You can deploy this model and start earning money today!

uriel/
taska-wav2vec-300m-max22-WF-epoch-16-batch-8-whisper-3e3

Name: taska-wav2vec-300m-max22-WF-epoch-16-batch-8-whisper-3e3
Availability: InStock

peft

tensorboard

safetensors

generated_from_trainer

openai/whisper-large-v2

adapter

Views

No views yet

taska-wav2vec-300m-max22-WF-epoch-16-batch-8-whisper-3e3

This model is a fine-tuned version of whisper-large-v2 on the uriel/audio_data_kaggle_train_taska dataset. It achieves the following results on the evaluation set:

Loss: 0.3716

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: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine_with_restarts
lr_scheduler_warmup_steps: 100
num_epochs: 4
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.8649	0.0888	250	0.7986
0.7419	0.1777	500	0.6707
0.6103	0.2665	750	0.6172
0.6134	0.3553	1000	0.5791
0.5575	0.4442	1250	0.5592
0.536	0.5330	1500	0.5423
0.4742	0.6218	1750	0.5245
0.5081	0.7107	2000	0.5082
0.5566	0.7995	2250	0.5061
0.5291	0.8883	2500	0.4951
0.5099	0.9772	2750	0.4870
0.4289	1.0657	3000	0.4779
0.4573	1.1546	3250	0.4696
0.4906	1.2434	3500	0.4652
0.4306	1.3322	3750	0.4594
0.4148	1.4211	4000	0.4546
0.4323	1.5099	4250	0.4473
0.4559	1.5987	4500	0.4416
0.4602	1.6876	4750	0.4374
0.4078	1.7764	5000	0.4293
0.4268	1.8652	5250	0.4257
0.4243	1.9541	5500	0.4211
0.3537	2.0426	5750	0.4141
0.3784	2.1315	6000	0.4110
0.3366	2.2203	6250	0.4081
0.3668	2.3091	6500	0.4051
0.3524	2.3980	6750	0.4002
0.336	2.4868	7000	0.3972
0.3622	2.5756	7250	0.3940
0.3674	2.6645	7500	0.3892
0.3591	2.7533	7750	0.3863
0.355	2.8421	8000	0.3824
0.3413	2.9310	8250	0.3791
0.3039	3.0195	8500	0.3794
0.3372	3.1084	8750	0.3788
0.3144	3.1972	9000	0.3776
0.2848	3.2860	9250	0.3769
0.3075	3.3749	9500	0.3749
0.2837	3.4637	9750	0.3742
0.3171	3.5525	10000	0.3729
0.2913	3.6414	10250	0.3723
0.3006	3.7302	10500	0.3720
0.287	3.8190	10750	0.3719
0.2878	3.9079	11000	0.3716
0.2781	3.9967	11250	0.3716

Framework versions

PEFT 0.15.2.dev0
Transformers 4.53.0.dev0
Pytorch 2.7.1+cu126
Datasets 3.6.0
Tokenizers 0.21.1

uriel/taska-wav2vec-300m-max22-WF-epoch-16-batch-8-whisper-3e3

taska-wav2vec-300m-max22-WF-epoch-16-batch-8-whisper-3e3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

uriel/
taska-wav2vec-300m-max22-WF-epoch-16-batch-8-whisper-3e3