aipib/Florence-2-FT-JP-DocVQA3

Florence-2-FT-JP-TF3

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: 1.8131

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-06
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• 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
2.6073	0.1115	100	2.8164
2.4221	0.2230	200	2.6756
2.2154	0.3344	300	2.5625
2.2387	0.4459	400	2.4688
2.2048	0.5574	500	2.3977
2.151	0.6689	600	2.3204
1.9731	0.7804	700	2.3002
2.0566	0.8919	800	2.2333
2.0783	1.0033	900	2.1996
1.8874	1.1148	1000	2.1861
1.721	1.2263	1100	2.1866
1.9372	1.3378	1200	2.1360
1.783	1.4493	1300	2.1260
1.7028	1.5608	1400	2.1044
1.7303	1.6722	1500	2.0908
1.727	1.7837	1600	2.0673
1.8337	1.8952	1700	2.0213
1.7386	2.0067	1800	2.0350
1.5987	2.1182	1900	2.0028
1.6715	2.2297	2000	2.0117
1.6533	2.3411	2100	2.0031
1.6372	2.4526	2200	1.9778
1.6107	2.5641	2300	1.9897
1.6603	2.6756	2400	1.9783
1.6026	2.7871	2500	1.9599
1.5174	2.8986	2600	1.9597
1.6123	3.0100	2700	1.9432
1.524	3.1215	2800	1.9508
1.494	3.2330	2900	1.9571
1.4581	3.3445	3000	1.9524
1.4949	3.4560	3100	1.9400
1.472	3.5674	3200	1.9243
1.5104	3.6789	3300	1.9217
1.4193	3.7904	3400	1.9148
1.3834	3.9019	3500	1.9052
1.4237	4.0134	3600	1.8973
1.3952	4.1249	3700	1.8927
1.3422	4.2363	3800	1.9020
1.4286	4.3478	3900	1.8967
1.3241	4.4593	4000	1.8944
1.443	4.5708	4100	1.8655
1.4617	4.6823	4200	1.8740
1.425	4.7938	4300	1.8686
1.4033	4.9052	4400	1.8816
1.3862	5.0167	4500	1.8716
1.3397	5.1282	4600	1.8668
1.2487	5.2397	4700	1.8623
1.2266	5.3512	4800	1.8447
1.3607	5.4627	4900	1.8524
1.2804	5.5741	5000	1.8594
1.4155	5.6856	5100	1.8503
1.3445	5.7971	5200	1.8643
1.2873	5.9086	5300	1.8506
1.3181	6.0201	5400	1.8384
1.363	6.1315	5500	1.8440
1.2563	6.2430	5600	1.8438
1.3474	6.3545	5700	1.8463
1.2846	6.4660	5800	1.8440
1.2657	6.5775	5900	1.8455
1.3529	6.6890	6000	1.8356
1.1427	6.8004	6100	1.8412
1.1655	6.9119	6200	1.8330
1.2756	7.0234	6300	1.8317
1.0908	7.1349	6400	1.8373
1.205	7.2464	6500	1.8288
1.1937	7.3579	6600	1.8277
1.2626	7.4693	6700	1.8218
1.2383	7.5808	6800	1.8202
1.2212	7.6923	6900	1.8199
1.2275	7.8038	7000	1.8142
1.2715	7.9153	7100	1.8170
1.2888	8.0268	7200	1.8172
1.1628	8.1382	7300	1.8200
1.239	8.2497	7400	1.8219
1.2209	8.3612	7500	1.8137
1.3384	8.4727	7600	1.8118
1.2151	8.5842	7700	1.8134
1.2329	8.6957	7800	1.8131
1.2236	8.8071	7900	1.8159
1.2664	8.9186	8000	1.8166
1.2349	9.0301	8100	1.8121
1.1064	9.1416	8200	1.8149
1.2612	9.2531	8300	1.8133
1.2315	9.3645	8400	1.8149
1.1879	9.4760	8500	1.8135
1.0139	9.5875	8600	1.8161
1.2237	9.6990	8700	1.8146
1.2138	9.8105	8800	1.8129
1.2941	9.9220	8900	1.8131

Framework versions

• Transformers 4.46.3
• Pytorch 2.3.0
• Datasets 3.2.0
• Tokenizers 0.20.3