README.md

metadata

license: mit
base_model: microsoft/git-base
tags:
  - generated_from_trainer
model-index:
  - name: GenerativeImage2Text-naruto
    results: []

GenerativeImage2Text-naruto

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: 0.0698
• Wer Score: 5.7759

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: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Score
7.3056	3.7037	50	4.4313	24.2069
2.1534	7.4074	100	0.3158	0.4310
0.0855	11.1111	150	0.0458	0.4655
0.0153	14.8148	200	0.0447	0.4655
0.0109	18.5185	250	0.0492	0.4483
0.0098	22.2222	300	0.0529	0.4483
0.0077	25.9259	350	0.0547	0.4828
0.0069	29.6296	400	0.0567	0.4483
0.0058	33.3333	450	0.0595	0.7414
0.0042	37.0370	500	0.0620	2.5517
0.0036	40.7407	550	0.0654	3.1379
0.0032	44.4444	600	0.0614	9.7414
0.0026	48.1481	650	0.0664	6.5517
0.001	51.8519	700	0.0670	7.4828
0.0006	55.5556	750	0.0662	7.5172
0.0006	59.2593	800	0.0670	8.7586
0.0003	62.9630	850	0.0678	7.7414
0.0003	66.6667	900	0.0685	6.7931
0.0002	70.3704	950	0.0688	6.3103
0.0002	74.0741	1000	0.0689	6.4828
0.0002	77.7778	1050	0.0691	6.1724
0.0002	81.4815	1100	0.0694	6.0862
0.0002	85.1852	1150	0.0695	6.1034
0.0002	88.8889	1200	0.0697	5.9828
0.0002	92.5926	1250	0.0698	5.8276
0.0002	96.2963	1300	0.0698	5.7759

Framework versions

• Transformers 4.40.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1