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: 0.0344
• Wer Score: 2.2807

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
7.2943	4.1667	50	4.4708	21.1278
2.2992	8.3333	100	0.4252	14.8722
0.1287	12.5	150	0.0311	0.6128
0.0161	16.6667	200	0.0280	2.4762
0.0049	20.8333	250	0.0304	2.4561
0.0022	25.0	300	0.0327	2.4085
0.0016	29.1667	350	0.0328	2.3333
0.0013	33.3333	400	0.0333	2.4298
0.0012	37.5	450	0.0341	2.3033
0.0011	41.6667	500	0.0344	2.2569
0.001	45.8333	550	0.0344	2.2744
0.001	50.0	600	0.0344	2.2807

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1