xtremedistil-l6-h384-uncased-v4.0

This model is a fine-tuned version of microsoft/xtremedistil-l6-h384-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5570
• F1 Macro: 0.6744
• F1 Micro: 0.6771
• Accuracy Balanced: 0.6742
• Accuracy: 0.6771
• Precision Macro: 0.6748
• Recall Macro: 0.6742
• Precision Micro: 0.6771
• Recall Micro: 0.6771

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: 40
• 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
0.5719	1.69	200	0.5779	0.6387	0.6559	0.6420	0.6559	0.6609	0.6420	0.6559	0.6559

eval result

Datasets	asadfgglie/nli-zh-tw-all/test	asadfgglie/BanBan_2024-10-17-facial_expressions-nli/test	eval_dataset	test_dataset
eval_loss	0.558	0.728	0.56	0.557
eval_f1_macro	0.676	0.494	0.682	0.674
eval_f1_micro	0.679	0.531	0.685	0.677
eval_accuracy_balanced	0.676	0.523	0.682	0.674
eval_accuracy	0.679	0.531	0.685	0.677
eval_precision_macro	0.676	0.53	0.682	0.675
eval_recall_macro	0.676	0.523	0.682	0.674
eval_precision_micro	0.679	0.531	0.685	0.677
eval_recall_micro	0.679	0.531	0.685	0.677
eval_runtime	9.08	0.195	1.746	7.023
eval_samples_per_second	936.093	4861.442	973.854	968.275
eval_steps_per_second	7.379	41.112	8.02	7.689
Size of dataset	8500	946	1700	6800

Framework versions

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