samchain/EconoDetect-US

EconoDetect

This model is a fine-tuned version of samchain/econo-sentence-v2 on the economics-relevance dataset. The base model is kept frozen during training, only the classification head is updated.

It achieves the following results on the evaluation set:

• Loss: 0.3973
• Accuracy: 0.8211
• F1: 0.7991
• Precision: 0.7895
• Recall: 0.8211

Model description

This model is designed to detect whether a text discusses topics related to the US economy.

Intended uses & limitations

The model can be used as a screening tool to remove texts that are not discussing US economy.

Training and evaluation data

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0005
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.5381	1.0	700	0.4333	0.7844	0.7894	0.7952	0.7844
0.4613	2.0	1400	0.4044	0.8328	0.7679	0.7856	0.8328
0.3523	3.0	2100	0.3973	0.8211	0.7991	0.7895	0.8211

Framework versions

• Transformers 4.50.0
• Pytorch 2.1.0+cu118
• Datasets 3.4.1
• Tokenizers 0.21.1