mohsinshah/git-base-dummy-3

git-base-500img-dataset

This model is a fine-tuned version of microsoft/git-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4161
• Wer Score: 2.0379

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Score
7.0698	3.23	50	4.5086	2.5298
2.6252	6.45	100	0.9823	2.2976
0.5497	9.68	150	0.4681	1.6707
0.2558	12.9	200	0.4162	1.7907
0.1551	16.13	250	0.4052	2.0984
0.1041	19.35	300	0.4054	2.0984
0.0764	22.58	350	0.4088	2.0576
0.0581	25.81	400	0.4054	2.0899
0.0462	29.03	450	0.4092	2.0484
0.0382	32.26	500	0.4118	2.1387
0.0329	35.48	550	0.4126	2.1315
0.0275	38.71	600	0.4139	2.0114
0.0255	41.94	650	0.4173	2.0098
0.0234	45.16	700	0.4155	2.0206
0.0226	48.39	750	0.4161	2.0379

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.3
• Tokenizers 0.13.3