xtremedistil-l6-h384-uncased-v3.0

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: 0.6934
• F1 Macro: 0.3279
• F1 Micro: 0.4782
• Accuracy Balanced: 0.4998
• Accuracy: 0.4782
• Precision Macro: 0.4890
• Recall Macro: 0.4998
• Precision Micro: 0.4782
• Recall Micro: 0.4782

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
0.6944	1.69	200	0.6934	0.3368	0.5079	0.5	0.5079	0.2540	0.5	0.5079	0.5079

eval result

Datasets	asadfgglie/nli-zh-tw-all/test	asadfgglie/BanBan_2024-10-17-facial_expressions-nli/test	eval_dataset	test_dataset
eval_loss	0.694	0.693	0.693	0.693
eval_f1_macro	0.298	0.33	0.337	0.328
eval_f1_micro	0.39	0.484	0.508	0.478
eval_accuracy_balanced	0.363	0.5	0.5	0.5
eval_accuracy	0.39	0.484	0.508	0.478
eval_precision_macro	0.269	0.492	0.254	0.489
eval_recall_macro	0.363	0.5	0.5	0.5
eval_precision_micro	0.39	0.484	0.508	0.478
eval_recall_micro	0.39	0.484	0.508	0.478
eval_runtime	9.197	0.194	0.059	0.174
eval_samples_per_second	924.224	4885.812	3188.956	4348.149
eval_steps_per_second	7.285	41.318	33.746	34.464
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