llama-3b-stance

This model is a fine-tuned version of meta-llama/Llama-3.2-3B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9008
• Accuracy: 0.5849
• Precision: 0.5752
• Recall: 0.5151
• F1: 0.5254

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	1.1089	0.4402	0.3888	0.3875	0.3878
No log	2.0	46	1.0486	0.4898	0.4587	0.4419	0.4417
No log	3.0	69	0.9954	0.5521	0.5380	0.4659	0.4701
No log	4.0	92	0.9417	0.5726	0.5441	0.4689	0.4756
No log	5.0	115	0.9531	0.5752	0.5827	0.4814	0.4831
No log	6.0	138	0.9205	0.5772	0.5606	0.4979	0.5068
No log	7.0	161	0.9181	0.5746	0.5624	0.5124	0.5207
No log	8.0	184	0.9157	0.5680	0.5498	0.5255	0.5298
No log	9.0	207	0.9032	0.5787	0.5601	0.5170	0.5258
No log	10.0	230	0.9008	0.5849	0.5752	0.5151	0.5254

Framework versions

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