Llama2_13B_Task2_semantic_pred

This model is a fine-tuned version of meta-llama/Llama-2-13b-hf on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3818
• Accuracy: 0.9087
• Precision: 0.9087
• Recall: 0.9087
• F1 score: 0.9087

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Accuracy	F1 score	Precision	Recall	Validation Loss
0.4564	0.2604	200	0.8214	0.8198	0.8311	0.8214	0.4378
0.3824	0.5208	400	0.8279	0.8279	0.8279	0.8279	0.4660
0.3609	0.7812	600	0.8631	0.8630	0.8635	0.8631	0.3303
0.3065	1.0417	800	0.8696	0.8695	0.8724	0.8696	0.3470
0.1987	1.3021	1000	0.8722	0.8722	0.8733	0.8722	0.3563
0.2043	1.5625	1200	0.9022	0.9020	0.9051	0.9022	0.3349
0.2193	1.8229	1400	0.8996	0.8996	0.8997	0.8996	0.3166
0.1674	2.0833	1600	0.8931	0.8930	0.8937	0.8931	0.3300
0.1226	2.3438	1800	0.3672	0.9087	0.9094	0.9087	0.9087
0.123	2.6042	2000	0.3862	0.9074	0.9091	0.9074	0.9073
0.0792	2.8646	2200	0.3818	0.9087	0.9087	0.9087	0.9087

Framework versions

• PEFT 0.11.1
• Transformers 4.44.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1