You can deploy this model and start earning money today!

nik1509/
xlsr_300m_telugu_lr_ablation_5e-5

Name: xlsr_300m_telugu_lr_ablation_5e-5
Availability: InStock

Views

No views yet

xlsr_300m_telugu_lr_ablation_5e-5

This model is a fine-tuned version of nik1509/xlsr300m_finetuned_telugu_basemodel on the None dataset. It achieves the following results on the evaluation set:

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
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_ratio: 0.1
num_epochs: 15
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
331.4871	0.6944	800	457.8265	0.3969	0.1656
300.1515	1.3889	1600	473.2798	0.3982	0.1648
302.2698	2.0833	2400	462.6637	0.4065	0.1681
316.8686	2.7778	3200	460.4791	0.4033	0.1665
293.3103	3.4722	4000	480.2869	0.4034	0.1659
267.0341	4.1667	4800	479.7465	0.4006	0.1660
257.2548	4.8611	5600	493.1853	0.3956	0.1652
278.7068	5.5556	6400	496.6875	0.3977	0.1659
249.3262	6.25	7200	501.6370	0.4008	0.1672
237.1323	6.9444	8000	515.2466	0.3974	0.1655
259.0331	7.6389	8800	514.8406	0.3994	0.1660
241.9655	8.3333	9600	519.2821	0.3981	0.1666
254.3616	9.0278	10400	523.8386	0.3994	0.1674
254.224	9.7222	11200	532.1626	0.3960	0.1668
236.7024	10.4167	12000	535.8199	0.3993	0.1672
230.1712	11.1111	12800	545.4752	0.3959	0.1669
188.1558	11.8056	13600	549.2825	0.3985	0.1672
246.9506	12.5	14400	555.6299	0.3959	0.1667
242.6181	13.1944	15200	550.8352	0.3986	0.1676
219.7372	13.8889	16000	548.5676	0.3994	0.1677
224.2553	14.5833	16800	543.6052	0.3996	0.1677