Hanhpt23/whisper-small-engmed-free_ED3-11

openai/whisper-small

This model is a fine-tuned version of openai/whisper-small on the pphuc25/EngMed dataset. It achieves the following results on the evaluation set:

• Loss: 0.0004
• Wer: 21.9876
• Cer: 17.0884

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: 0.0001
• train_batch_size: 8
• 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: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.7291	1.0	386	0.3415	25.0142	20.1560
0.472	2.0	772	0.2196	27.4412	24.5293
0.2716	3.0	1158	0.1239	31.2274	28.4146
0.1409	4.0	1544	0.0740	38.1165	34.4281
0.1283	5.0	1930	0.0465	40.1171	35.9613
0.0631	6.0	2316	0.0310	36.5868	30.1992
0.0582	7.0	2702	0.0177	32.1775	26.2129
0.0531	8.0	3088	0.0131	31.8391	27.8444
0.0238	9.0	3474	0.0091	24.2508	18.5448
0.0149	10.0	3860	0.0060	24.8696	19.3220
0.0057	11.0	4246	0.0050	26.8193	21.7715
0.0077	12.0	4632	0.0031	23.0677	19.1910
0.0073	13.0	5018	0.0028	24.7584	19.3135
0.0052	14.0	5404	0.0014	25.8657	19.1331
0.0031	15.0	5790	0.0009	21.5274	17.0940
0.0039	16.0	6176	0.0007	22.1520	17.1637
0.0013	17.0	6562	0.0006	22.9021	17.7620
0.0009	18.0	6948	0.0005	21.9899	17.3717
0.001	19.0	7334	0.0004	22.1923	17.4027
0.0004	20.0	7720	0.0004	21.9876	17.0884

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.19.1