kugler/gbert-large-AmDi.tiny-synset-classifier

gbert_synset_classifier_amdi_tiny

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

• Loss: 0.6469
• Accuracy: 0.8376
• F1: 0.8366
• Precision: 0.8446
• Recall: 0.8376
• F1 Macro: 0.8202
• Precision Macro: 0.7900
• Recall Macro: 0.8625
• F1 Micro: 0.8376
• Precision Micro: 0.8376
• Recall Micro: 0.8376

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	F1 Macro	Precision Macro	Recall Macro	F1 Micro	Precision Micro	Recall Micro
2.7014	0.8584	100	1.1834	0.6919	0.6499	0.6573	0.6919	0.4964	0.5670	0.4997	0.6919	0.6919	0.6919
0.7696	1.7167	200	0.5970	0.8241	0.8188	0.8227	0.8241	0.7784	0.7603	0.8062	0.8241	0.8241	0.8241
0.4739	2.5751	300	0.5408	0.8321	0.8290	0.8381	0.8321	0.8082	0.7815	0.8508	0.8321	0.8321	0.8321
0.3743	3.4335	400	0.5343	0.8426	0.8385	0.8480	0.8426	0.8269	0.8046	0.8598	0.8426	0.8426	0.8426
0.2931	4.2918	500	0.5188	0.8469	0.8465	0.8516	0.8469	0.8312	0.8133	0.8552	0.8469	0.8469	0.8469
0.2175	5.1502	600	0.5697	0.8426	0.8419	0.8506	0.8426	0.8295	0.8093	0.8572	0.8426	0.8426	0.8426
0.1689	6.0086	700	0.5781	0.8426	0.8421	0.8470	0.8426	0.8322	0.8164	0.8540	0.8426	0.8426	0.8426
0.1174	6.8670	800	0.6469	0.8376	0.8366	0.8446	0.8376	0.8202	0.7900	0.8625	0.8376	0.8376	0.8376

Framework versions

• Transformers 4.45.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.20.3