subh_w2lm_base_distill_noisy_teacher_mozilla_epochs_50_batch_16

This model is a fine-tuned version of rohitp1/ws_w2lm_base_plus_finetune_teacher_noise_mozilla_100_epochs_batch_8 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4070
• Wer: 0.3226

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 256
• total_train_batch_size: 4096
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1283	7.31	250	0.3295	0.3266
0.1111	14.63	500	0.3444	0.3236
0.0939	21.94	750	0.3664	0.3236
0.0826	29.26	1000	0.3828	0.3224
0.0751	36.57	1250	0.3977	0.3223
0.0703	43.89	1500	0.4070	0.3226

Framework versions

• Transformers 4.29.2
• Pytorch 1.12.1
• Datasets 2.8.0
• Tokenizers 0.13.2