pkemon_cap_v0

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

• Loss: 7.6491
• Wer Score: 127.2727

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Score
11.2497	0.17	2	10.0191	96.6364
9.9157	0.35	4	9.5544	111.1818
9.4907	0.52	6	9.1167	143.5909
9.0975	0.7	8	8.8422	154.5455
8.8568	0.87	10	8.6143	144.6364
8.6299	1.04	12	8.4336	118.7727
8.4659	1.22	14	8.2808	112.4091
8.3233	1.39	16	8.1538	124.3636
8.2213	1.57	18	8.0420	122.8636
8.0876	1.74	20	7.9463	124.5
7.9863	1.91	22	7.8647	153.9545
7.9169	2.09	24	7.7966	156.0
7.8652	2.26	26	7.7400	155.5455
7.8245	2.43	28	7.6962	142.0909
7.7512	2.61	30	7.6659	129.9545
7.7344	2.78	32	7.6491	127.2727

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.0+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3