Chenchuhui/Icelandic-finetuned-data-augmentation

Icelandic_model_aug_checkpoints

This model is a fine-tuned version of carlosdanielhernandezmena/wav2vec2-large-xlsr-53-icelandic-ep10-1000h on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2766
• Wer: 0.2136

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.2	10	0.6818	0.3244
No log	0.4	20	0.4570	0.2841
No log	0.6	30	0.4281	0.2562
No log	0.8	40	0.4284	0.2282
0.8009	1.0	50	0.4335	0.2260
0.8009	1.2	60	0.4302	0.2159
0.8009	1.4	70	0.4113	0.2114
0.8009	1.6	80	0.3847	0.2159
0.8009	1.8	90	0.3795	0.2148
0.4567	2.0	100	0.3716	0.2058
0.4567	2.2	110	0.3848	0.2081
0.4567	2.4	120	0.3797	0.2036
0.4567	2.6	130	0.3695	0.2036
0.4567	2.8	140	0.3545	0.1969
0.3723	3.0	150	0.3433	0.1957
0.3723	3.2	160	0.3213	0.1957
0.3723	3.4	170	0.3068	0.1913
0.3723	3.6	180	0.3049	0.1902
0.3723	3.8	190	0.3039	0.1991
0.3259	4.0	200	0.2986	0.1913
0.3259	4.2	210	0.3181	0.1969
0.3259	4.4	220	0.3216	0.1857
0.3259	4.6	230	0.3336	0.1913
0.3259	4.8	240	0.3317	0.1980
0.2809	5.0	250	0.2948	0.2025
0.2809	5.2	260	0.3165	0.1879
0.2809	5.4	270	0.3289	0.2081
0.2809	5.6	280	0.3071	0.1913
0.2809	5.8	290	0.3116	0.1957
0.2711	6.0	300	0.2441	0.2136
0.2711	6.2	310	0.2961	0.2013
0.2711	6.4	320	0.3473	0.2069
0.2711	6.6	330	0.2201	0.2058
0.2711	6.8	340	0.2004	0.1991
0.2578	7.0	350	0.2213	0.2081
0.2578	7.2	360	0.3698	0.2081
0.2578	7.4	370	0.3921	0.2114
0.2578	7.6	380	0.3861	0.2204
0.2578	7.8	390	0.3792	0.2271
0.2503	8.0	400	0.3084	0.2092
0.2503	8.2	410	0.3273	0.2058
0.2503	8.4	420	0.3965	0.2069
0.2503	8.6	430	0.2838	0.2036
0.2503	8.8	440	0.3243	0.2148
0.2485	9.0	450	0.3139	0.2136
0.2485	9.2	460	0.3155	0.2036
0.2485	9.4	470	0.3140	0.2103
0.2485	9.6	480	0.2902	0.2349
0.2485	9.8	490	0.2389	0.2047
0.2567	10.0	500	0.2766	0.2136

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0