fine-turn-ColQwen2-vn

This model is a fine-tuned version of vidore/colqwen2.5-base on the quyet498/train-col-vn dataset. It achieves the following results on the evaluation set:

• Loss: 0.0949
• Model Preparation Time: 0.0089

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 1.2

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time
No log	0.0002	1	0.9128	0.0089
0.2186	0.0444	200	0.3293	0.0089
0.2364	0.0889	400	0.2411	0.0089
0.2301	0.1333	600	0.1937	0.0089
0.2082	0.1778	800	0.1852	0.0089
0.1739	0.2222	1000	0.1614	0.0089
0.1597	0.2667	1200	0.1650	0.0089
0.2538	0.3111	1400	0.1396	0.0089
0.1697	0.3556	1600	0.1417	0.0089
0.1634	0.4	1800	0.1236	0.0089
0.2365	0.4444	2000	0.1382	0.0089
0.1706	0.4889	2200	0.1289	0.0089
0.1404	0.5333	2400	0.1173	0.0089
0.127	0.5778	2600	0.1263	0.0089
0.1721	0.6222	2800	0.1131	0.0089
0.2178	0.6667	3000	0.1112	0.0089
0.1041	0.7111	3200	0.1104	0.0089
0.1538	0.7556	3400	0.1042	0.0089
0.1245	0.8	3600	0.1095	0.0089
0.1164	0.8444	3800	0.1053	0.0089
0.0894	0.8889	4000	0.1070	0.0089
0.1224	0.9333	4200	0.1061	0.0089
0.1275	0.9778	4400	0.0999	0.0089
0.1214	1.0222	4600	0.0993	0.0089
0.0985	1.0667	4800	0.0970	0.0089
0.0948	1.1111	5000	0.0953	0.0089
0.0602	1.1556	5200	0.0957	0.0089
0.0867	1.2	5400	0.0949	0.0089

Framework versions

• Transformers 4.53.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.4