gpt2-sentiment_analysis

This model is a fine-tuned version of openai-community/gpt2 on the financial_phrasebank dataset. It achieves the following results on the evaluation set:

• Loss: 0.6571
• Accuracy: {'accuracy': 0.8239339752407153}

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: 0.0006
• train_batch_size: 4
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9981	257	0.4654	{'accuracy': 0.8239339752407153}
0.6288	2.0	515	0.4266	{'accuracy': 0.8266850068775791}
0.6288	2.9981	772	0.4558	{'accuracy': 0.8225584594222833}
0.3201	4.0	1030	0.4550	{'accuracy': 0.811554332874828}
0.3201	4.9981	1287	0.4223	{'accuracy': 0.8294360385144429}
0.2464	6.0	1545	0.4637	{'accuracy': 0.8335625859697386}
0.2464	6.9981	1802	0.5243	{'accuracy': 0.8184319119669876}
0.1859	8.0	2060	0.5482	{'accuracy': 0.8335625859697386}
0.1859	8.9981	2317	0.6443	{'accuracy': 0.8335625859697386}
0.1381	9.9806	2570	0.6571	{'accuracy': 0.8239339752407153}

Framework versions

• PEFT 0.11.1
• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1