sentiment_pc_under_sampler

This model is a fine-tuned version of ahmedrachid/FinancialBERT-Sentiment-Analysis on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5934
• Accuracy: 0.8235
• F1: 0.8236

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
• 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
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.4464	50	0.5398	0.7797	0.7809
No log	0.8929	100	0.4638	0.8203	0.8221
No log	1.3393	150	0.4749	0.8183	0.8196
No log	1.7857	200	0.4944	0.8144	0.8150
No log	2.2321	250	0.5050	0.8157	0.8158
No log	2.6786	300	0.5470	0.8085	0.8079
No log	3.125	350	0.5299	0.8196	0.8200
No log	3.5714	400	0.5651	0.8150	0.8146
No log	4.0179	450	0.5684	0.8288	0.8294
0.3419	4.4643	500	0.5934	0.8235	0.8236
0.3419	4.9107	550	0.5976	0.8203	0.8208

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1