table-transformer-deep-dust-69

This model is a fine-tuned version of microsoft/table-transformer-detection on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7515
• Loss Ce: 0.0037
• Loss Bbox: 0.1065
• Cardinality Error: 1.0018
• Giou: 89.1969

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Loss Ce	Loss Bbox	Cardinality Error	Giou
1.0999	0.2	500	1.1217	0.1327	0.1393	1.0018	85.2782
0.7151	0.4	1000	0.9541	0.0420	0.1291	1.0018	86.5834
0.5844	0.6	1500	0.8440	0.0255	0.1161	1.0018	88.1012
0.5046	0.8	2000	0.7986	0.0177	0.1109	1.0018	88.6059
0.4826	1.0	2500	0.7727	0.0132	0.1078	1.0018	88.9256
0.3942	1.2	3000	0.8233	0.0097	0.1159	1.0018	88.2610
0.4361	1.4	3500	0.7496	0.0080	0.1054	1.0018	89.2268
0.3225	1.6	4000	0.7584	0.0066	0.1067	1.0018	89.0580
0.371	1.8	4500	0.7559	0.0060	0.1068	1.0018	89.1710
0.3759	2.0	5000	0.7732	0.0051	0.1094	1.0018	88.8865
0.3031	2.2	5500	0.7353	0.0049	0.1040	1.0018	89.4293
0.3806	2.4	6000	0.7551	0.0042	0.1068	1.0018	89.1146
0.2889	2.6	6500	0.7529	0.0039	0.1066	1.0018	89.1589
0.2847	2.8	7000	0.7487	0.0038	0.1060	1.0018	89.2153
0.2906	3.0	7500	0.7515	0.0037	0.1065	1.0018	89.1969

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3