eolang/sw-peft-2

This model is a fine-tuned version of openai/whisper-medium on the Common Voice 19 dataset. It achieves the following results on the evaluation set:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 0.001
train_batch_size: 8
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: 50
num_epochs: 1
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
1.5084	0.0191	100	1.0225
0.9199	0.0382	200	0.9041
0.8295	0.0573	300	0.8428
0.8404	0.0764	400	0.8142
0.8215	0.0955	500	0.8141
0.7527	0.1147	600	0.7640
0.7494	0.1338	700	0.7652
0.7693	0.1529	800	0.7081
0.7104	0.1720	900	0.7238
0.6975	0.1911	1000	0.6990
0.7164	0.2102	1100	0.7002
0.6693	0.2293	1200	0.6842
0.7043	0.2484	1300	0.6831
0.6521	0.2675	1400	0.6527
0.6468	0.2866	1500	0.6563
0.6376	0.3058	1600	0.6180
0.6008	0.3249	1700	0.6223
0.6353	0.3440	1800	0.6113
0.5885	0.3631	1900	0.6033
0.598	0.3822	2000	0.5987
0.5749	0.4013	2100	0.5792
0.5714	0.4204	2200	0.5772
0.5438	0.4395	2300	0.5688
0.5442	0.4586	2400	0.5711
0.5165	0.4777	2500	0.5588
0.4971	0.4968	2600	0.5408
0.5026	0.5160	2700	0.5365
0.5278	0.5351	2800	0.5112
0.5371	0.5542	2900	0.5160
0.5013	0.5733	3000	0.5041
0.4867	0.5924	3100	0.4978
0.4938	0.6115	3200	0.4830
0.4522	0.6306	3300	0.4798
0.4515	0.6497	3400	0.4751
0.4593	0.6688	3500	0.4631
0.4539	0.6879	3600	0.4561
0.4557	0.7071	3700	0.4467
0.417	0.7262	3800	0.4419
0.4251	0.7453	3900	0.4368
0.4062	0.7644	4000	0.4277
0.3815	0.7835	4100	0.4271
0.3832	0.8026	4200	0.4155
0.3818	0.8217	4300	0.4098
0.3988	0.8408	4400	0.4005
0.4073	0.8599	4500	0.3964
0.3894	0.8790	4600	0.3898
0.3464	0.8981	4700	0.3858
0.3626	0.9173	4800	0.3800
0.3753	0.9364	4900	0.3771
0.3734	0.9555	5000	0.3733
0.3362	0.9746	5100	0.3718
0.3607	0.9937	5200	0.3697