README.md

metadata

library_name: transformers
base_model: studio-ousia/luke-base-lite
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: 1ccfffeacad1f0b9e6353344b490bd38
    results: []

1ccfffeacad1f0b9e6353344b490bd38

This model is a fine-tuned version of studio-ousia/luke-base-lite on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 0.1864
• Data Size: 1.0
• Epoch Runtime: 45.3066
• Accuracy: 0.9279
• F1 Macro: 0.8799

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	1.8419	0	2.2204	0.0338	0.0128
No log	1	500	1.6823	0.0078	2.8052	0.2908	0.0751
No log	2	1000	1.5591	0.0156	3.1628	0.3488	0.0862
No log	3	1500	1.1409	0.0312	3.9978	0.5736	0.2364
No log	4	2000	0.7112	0.0625	5.6075	0.7601	0.6169
0.0506	5	2500	0.4577	0.125	8.4992	0.8387	0.7905
0.344	6	3000	0.2268	0.25	13.9797	0.9189	0.8742
0.0383	7	3500	0.2045	0.5	24.5039	0.9254	0.8853
0.1978	8.0	4000	0.1479	1.0	46.4312	0.9304	0.8919
0.1522	9.0	4500	0.1909	1.0	45.6419	0.9279	0.8888
0.1482	10.0	5000	0.2014	1.0	45.1532	0.9264	0.8917
0.1233	11.0	5500	0.1883	1.0	44.6967	0.9294	0.8932
0.1042	12.0	6000	0.1864	1.0	45.3066	0.9279	0.8799

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1