_base_nougat_logs

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

• Loss: 0.4360

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 6
• total_train_batch_size: 48
• 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss
2.0261	0.9901	83	1.9705
1.8445	1.9920	167	1.7617
1.6923	2.9940	251	1.6539
1.597	3.9960	335	1.5871
1.5287	4.9980	419	1.5196
1.46	6.0	503	1.4804
1.3647	6.9901	586	1.4360
1.289	7.9920	670	1.3772
1.1741	8.9940	754	1.2350
0.9947	9.9960	838	1.0415
0.7889	10.9980	922	0.9238
0.6771	12.0	1006	0.7884
0.6256	12.9901	1089	0.6646
0.5402	13.9920	1173	0.6095
0.5252	14.9940	1257	0.5702
0.441	15.9960	1341	0.5282
0.4077	16.9980	1425	0.5030
0.3841	18.0	1509	0.4855
0.3762	18.9901	1592	0.4703
0.3611	19.9920	1676	0.4587
0.3486	20.9940	1760	0.4486
0.3679	21.9960	1844	0.4416
0.3356	22.9980	1928	0.4400
0.3343	24.0	2012	0.4387
0.3229	24.9901	2095	0.4410
0.2928	25.9920	2179	0.4377
0.3042	26.9940	2263	0.4393
0.3439	27.9960	2347	0.4353
0.3286	28.9980	2431	0.4365
0.353	29.7018	2490	0.4360

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.20.3