distilroberta-finetuned-financial-news-sentiment-analysis-clickbait-task1-20-epoch-post

This model is a fine-tuned version of mrm8488/distilroberta-finetuned-financial-news-sentiment-analysis on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.7757
• Accuracy: 0.6775

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	200	0.8214	0.6175
No log	2.0	400	0.7636	0.675
0.8421	3.0	600	0.7821	0.705
0.8421	4.0	800	0.8773	0.68
0.4325	5.0	1000	0.9866	0.6775
0.4325	6.0	1200	1.1958	0.6825
0.4325	7.0	1400	1.3580	0.6775
0.1921	8.0	1600	1.7104	0.7
0.1921	9.0	1800	1.9208	0.69
0.099	10.0	2000	2.1611	0.6675
0.099	11.0	2200	2.2038	0.68
0.099	12.0	2400	2.3109	0.68
0.0475	13.0	2600	2.4423	0.69
0.0475	14.0	2800	2.5835	0.6825
0.0301	15.0	3000	2.6523	0.6625
0.0301	16.0	3200	2.6894	0.675
0.0301	17.0	3400	2.6711	0.6875
0.0171	18.0	3600	2.7655	0.6725
0.0171	19.0	3800	2.7620	0.68
0.0125	20.0	4000	2.7757	0.6775

Framework versions

• Transformers 4.44.0.dev0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1