visobert_v1

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

• Loss: 0.4766
• Accuracy: 0.9337
• Precision Macro: 0.8527
• Recall Macro: 0.8055
• F1 Macro: 0.8251
• F1 Weighted: 0.9316

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: 3e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• 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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision Macro	Recall Macro	F1 Macro	F1 Weighted
0.377	1.0	90	0.2037	0.9406	0.9012	0.7694	0.8068	0.9350
0.1659	2.0	180	0.2094	0.9356	0.8396	0.8232	0.8309	0.9348
0.0966	3.0	270	0.2278	0.9381	0.8463	0.8165	0.8298	0.9367
0.0696	4.0	360	0.2619	0.9318	0.8438	0.7756	0.8003	0.9280
0.0468	5.0	450	0.3120	0.9324	0.8362	0.8128	0.8234	0.9313
0.0337	6.0	540	0.3576	0.9311	0.8376	0.7912	0.8103	0.9287
0.0244	7.0	630	0.3796	0.9292	0.8428	0.7816	0.8051	0.9261
0.019	8.0	720	0.4309	0.9349	0.8612	0.8070	0.8286	0.9327
0.0094	9.0	810	0.4022	0.9337	0.8565	0.8134	0.8318	0.9319
0.0098	10.0	900	0.4181	0.9349	0.8534	0.8062	0.8259	0.9329
0.0039	11.0	990	0.4484	0.9330	0.8542	0.8091	0.8281	0.9311
0.0028	12.0	1080	0.4580	0.9349	0.8554	0.8106	0.8294	0.9330
0.0028	13.0	1170	0.4554	0.9318	0.8613	0.7998	0.8242	0.9292
0.0031	14.0	1260	0.4575	0.9330	0.8579	0.8009	0.8237	0.9306
0.0018	15.0	1350	0.4547	0.9356	0.8617	0.8068	0.8291	0.9333
0.0004	16.0	1440	0.4631	0.9343	0.8455	0.8182	0.8305	0.9331
0.0006	17.0	1530	0.4642	0.9356	0.8542	0.8152	0.8319	0.9339
0.0008	18.0	1620	0.4736	0.9343	0.8534	0.8141	0.8311	0.9326
0.0014	19.0	1710	0.4753	0.9337	0.8527	0.8055	0.8251	0.9316
0.0009	20.0	1800	0.4766	0.9337	0.8527	0.8055	0.8251	0.9316

Framework versions

• Transformers 4.55.0
• Pytorch 2.7.0+cu126
• Datasets 4.0.0
• Tokenizers 0.21.4