Finetuned Whisper large-v3-turbo for darija speech translation

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the Darija-C 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: 16
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: 100
training_steps: 1000
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Bleu
2.4688	2.2727	50	2.1049	0.0694
0.8484	4.5455	100	0.9947	0.1871
0.3358	6.8182	150	0.2579	0.5827
0.1395	9.0909	200	0.0936	0.6840
0.0669	11.3636	250	0.0383	0.7778
0.0421	13.6364	300	0.0214	0.7793
0.0293	15.9091	350	0.0195	0.8053
0.0228	18.1818	400	0.0102	0.8019
0.0132	20.4545	450	0.0065	0.8014
0.011	22.7273	500	0.0050	0.8053
0.0094	25.0	550	0.0025	0.8080
0.0048	27.2727	600	0.0009	0.8080
0.002	29.5455	650	0.0007	0.8080
0.0011	31.8182	700	0.0005	0.8080
0.0007	34.0909	750	0.0005	0.8080
0.0006	36.3636	800	0.0004	0.8080
0.0004	38.6364	850	0.0004	0.8080
0.0003	40.9091	900	0.0004	0.8080
0.0002	43.1818	950	0.0004	0.8080
0.0002	45.4545	1000	0.0004	0.8080