Tuan-Bao/florence-2

donut-finetuned-v2.1

This model is a fine-tuned version of naver-clova-ix/donut-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0396

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-05
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
4.2166	0.0394	200	2.8622
1.7778	0.0788	400	1.6184
1.3467	0.1182	600	1.0477
0.6550	0.1577	800	0.6482
0.5164	0.1971	1000	0.4542
0.3735	0.2365	1200	0.3639
0.4138	0.2759	1400	0.2679
0.2386	0.3153	1600	0.2156
0.2029	0.3547	1800	0.1628
0.2021	0.3942	2000	0.1289
0.1504	0.4336	2200	0.1080
0.1357	0.4730	2400	0.0904
0.0719	0.5124	2600	0.0798
0.0880	0.5518	2800	0.0715
0.0867	0.5912	3000	0.0616
0.0563	0.6307	3200	0.0601
0.1239	0.6701	3400	0.0550
0.0483	0.7095	3600	0.0513
0.0445	0.7489	3800	0.0480
0.0259	0.7883	4000	0.0469
0.0223	0.8277	4200	0.0448
0.0665	0.8672	4400	0.0426
0.0254	0.9066	4600	0.0413
0.0205	0.9460	4800	0.0404
0.0353	0.9854	5000	0.0396

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2