llama-7b-stance

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0488
• Accuracy: 0.5583
• Precision: 0.5489
• Recall: 0.5339
• F1: 0.5316

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: 0.0002
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	1.0	23	1.6614	0.3502	0.3641	0.3722	0.3428
No log	2.0	46	1.3645	0.4305	0.4308	0.4285	0.4071
No log	3.0	69	1.1866	0.4893	0.4687	0.4613	0.4567
No log	4.0	92	1.1019	0.5343	0.5079	0.4942	0.4964
No log	5.0	115	1.0842	0.5516	0.5335	0.4926	0.4995
No log	6.0	138	1.0671	0.5634	0.5589	0.5165	0.5210
No log	7.0	161	1.0930	0.5435	0.5430	0.5265	0.5195
No log	8.0	184	1.0652	0.5440	0.5324	0.5368	0.5260
No log	9.0	207	1.0162	0.5619	0.5352	0.5279	0.5295
No log	10.0	230	1.0488	0.5583	0.5489	0.5339	0.5316

Framework versions

• PEFT 0.14.0
• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0