git-base-pokemon

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: 0.0427
• Wer Score: 1.6214

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Score
7.2817	4.17	50	4.4600	19.8362
2.2912	8.33	100	0.4293	9.7913
0.1282	12.5	150	0.0391	0.3738
0.0156	16.67	200	0.0363	3.2961
0.0046	20.83	250	0.0396	1.9320
0.0022	25.0	300	0.0405	1.7464
0.0016	29.17	350	0.0414	1.9211
0.0013	33.33	400	0.0419	1.5837
0.0012	37.5	450	0.0424	1.7609
0.0011	41.67	500	0.0425	1.6614
0.001	45.83	550	0.0427	1.625
0.001	50.0	600	0.0427	1.6214

Framework versions

• Transformers 4.37.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.16.0
• Tokenizers 0.15.0