Florence-2-FT-JP-DocVQA

This model is a fine-tuned version of microsoft/Florence-2-base-ft on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.4379

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: 2e-06
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
12.6548	0.9662	100	2.8146
11.1776	1.9275	200	2.6563
10.6997	2.8889	300	2.6233
10.4599	3.8502	400	2.5377
10.1308	4.8116	500	2.5149
10.0422	5.7729	600	2.4952
9.9102	6.7343	700	2.4699
9.9127	7.6957	800	2.4560
9.8393	8.6570	900	2.4552
9.7839	9.6184	1000	2.4379

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0