whisper-small-lug-4src

This model is a fine-tuned version of sulaimank/whisper-small-CV-Fleurs-lg-300 on the None 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: 1e-05
train_batch_size: 8
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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
num_epochs: 30

Training Loss	Epoch	Step	Cer	Validation Loss	Wer
0.6768	1.1508	1000	0.2882	0.4067	0.3768
0.3794	2.3017	2000	0.2152	0.3882	0.3322
0.3018	3.4525	3000	0.4216	0.2716	0.1417
0.2216	4.6033	4000	0.4807	0.2945	0.1581
0.1689	5.7542	5000	0.5392	0.2358	0.1309
0.0748	6.9050	6000	0.5808	0.2331	0.1226
0.0453	8.0553	7000	0.6295	0.2382	0.1290
0.0276	9.2061	8000	0.6217	0.2413	0.1291
0.0166	10.3569	9000	0.6351	0.2300	0.1183
0.0115	11.5078	10000	0.6329	0.2366	0.1237
0.0096	12.6586	11000	0.6458	0.2428	0.1259
0.0059	13.8094	12000	0.6414	0.2323	0.1182
0.0045	14.9603	13000	0.6418	0.2638	0.1463
0.0043	16.1105	14000	0.6474	0.2362	0.1269
0.0026	17.2614	15000	0.6542	0.2284	0.1259
0.0026	18.4122	16000	0.6498	0.2253	0.1164
0.0013	19.5630	17000	0.6504	0.2758	0.1572
0.0015	20.7139	18000	0.6608	0.2187	0.1161
0.0015	21.8647	19000	0.6602	0.2234	0.1173
0.0015	23.0150	20000	0.6661	0.2455	0.1296
0.0008	24.1658	21000	0.6625	0.2265	0.1214
0.0003	25.3166	22000	0.6698	0.2498	0.1384
0.0003	26.4675	23000	0.6725	0.2347	0.1234
0.0001	27.6183	24000	0.6738	0.2187	0.1194
0.0001	28.7691	25000	0.6766	0.2133	0.1144
0.0001	29.9200	26000	0.6772	0.2145	0.1136