mDeBERTa-v3-base-xnli-multilingual-zeroshot-v3.0-only-non-nli

This model is a fine-tuned version of MoritzLaurer/mDeBERTa-v3-base-xnli-multilingual-nli-2mil7 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2718
• F1 Macro: 0.9088
• F1 Micro: 0.9089
• Accuracy Balanced: 0.9089
• Accuracy: 0.9089
• Precision Macro: 0.9092
• Recall Macro: 0.9089
• Precision Micro: 0.9089
• Recall Micro: 0.9089

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: 16
• eval_batch_size: 128
• seed: 20241201
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Macro	F1 Micro	Accuracy Balanced	Accuracy	Precision Macro	Recall Macro	Precision Micro	Recall Micro
0.2798	1.69	200	0.3328	0.8677	0.8677	0.8681	0.8677	0.8678	0.8681	0.8677	0.8677

Eval results

Datasets	asadfgglie/nli-zh-tw-all/test	asadfgglie/BanBan_2024-10-17-facial_expressions-nli/test	eval_dataset	test_dataset
eval_loss	0.667	0.294	0.381	0.272
eval_f1_macro	0.711	0.901	0.868	0.909
eval_f1_micro	0.713	0.901	0.868	0.909
eval_accuracy_balanced	0.71	0.901	0.867	0.909
eval_accuracy	0.713	0.901	0.868	0.909
eval_precision_macro	0.711	0.901	0.868	0.909
eval_recall_macro	0.71	0.901	0.867	0.909
eval_precision_micro	0.713	0.901	0.868	0.909
eval_recall_micro	0.713	0.901	0.868	0.909
eval_runtime	568.387	4.571	0.829	3.382
eval_samples_per_second	14.955	206.945	227.909	223.805
eval_steps_per_second	0.118	1.75	2.412	1.774
epoch	2.99	2.99	2.99	2.99
Size of dataset	8500	946	189	757

Framework versions

• Transformers 4.33.3
• Pytorch 2.5.1+cu121
• Datasets 2.14.7
• Tokenizers 0.13.3