xbilek25/whisper-small-train-basic_1000_v1.0

basic_train_basic_test 1000 similar params: per_device_train_batch_size=16, # bylo 16 a pod tim 1 gradient_accumulation_steps=1, warmup_steps=300, max_steps=3000

This model is a fine-tuned version of openai/whisper-small on the xbilek25/basic_train_set_en_last3cs_1000 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1515
• Wer: 15.2619

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: 1e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 300
• training_steps: 3000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0008	9.01	600	0.1497	17.2540
0.0002	19.0	1200	0.1470	16.1723
0.0001	28.01	1800	0.1501	16.1625
0.0001	38.0	2400	0.1509	14.5277
0.0001	47.01	3000	0.1515	15.2619

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.15.2