llama-8b-wiki10-31k-adapt-multilabel-classify

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:

• F1 Micro: 0.0057
• F1 Macro: 0.0005
• Precision At 5: 0.0146
• Recall At 5: 0.0038
• Psp At 5: 0.0048
• Precision At 8: 0.0099
• Recall At 8: 0.0041
• Psp At 8: 0.0040
• Precision At 15: 0.0062
• Recall At 15: 0.0047
• Psp At 15: 0.0037
• Precision At 25: 0.0043
• Recall At 25: 0.0055
• Psp At 25: 0.0039
• Rare F1 Micro: 0.0
• Rare F1 Macro: 0.0
• Rare Precision: 0.0
• Rare Recall: 0.0
• Rare Precision At 5: 0.0001
• Rare Recall At 5: 0.0004
• Rare Precision At 8: 0.0001
• Rare Recall At 8: 0.0005
• Rare Precision At 15: 0.0001
• Rare Recall At 15: 0.0006
• Rare Precision At 25: 0.0001
• Rare Recall At 25: 0.0008
• Not Rare F1 Micro: 0.0060
• Not Rare F1 Macro: 0.0008
• Not Rare Precision: 0.5636
• Not Rare Recall: 0.0030
• Not Rare Precision At 5: 0.0147
• Not Rare Recall At 5: 0.0040
• Not Rare Precision At 8: 0.0102
• Not Rare Recall At 8: 0.0044
• Not Rare Precision At 15: 0.0059
• Not Rare Recall At 15: 0.0048
• Not Rare Precision At 25: 0.0037
• Not Rare Recall At 25: 0.0050
• Loss: -3.3437

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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: 5
• num_epochs: 3

Training results

Training Loss	Epoch	Step	F1 Micro	F1 Macro	Precision At 5	Recall At 5	Psp At 5	Precision At 8	Recall At 8	Psp At 8	Precision At 15	Recall At 15	Psp At 15	Precision At 25	Recall At 25	Psp At 25	Rare F1 Micro	Rare F1 Macro	Rare Precision	Rare Recall	Rare Precision At 5	Rare Recall At 5	Rare Precision At 8	Rare Recall At 8	Rare Precision At 15	Rare Recall At 15	Rare Precision At 25	Rare Recall At 25	Not Rare F1 Micro	Not Rare F1 Macro	Not Rare Precision	Not Rare Recall	Not Rare Precision At 5	Not Rare Recall At 5	Not Rare Precision At 8	Not Rare Recall At 8	Not Rare Precision At 15	Not Rare Recall At 15	Not Rare Precision At 25	Not Rare Recall At 25	Validation Loss
-3.5077	1.0	10850	0.1363	0.0081	0.4242	0.1163	0.1328	0.3148	0.1359	0.1194	0.1935	0.1539	0.1045	0.1237	0.1625	0.0986	0.0038	0.0010	0.0348	0.0020	0.0028	0.0073	0.0018	0.0076	0.0010	0.0079	0.0006	0.0083	0.1439	0.0129	0.5919	0.0819	0.4252	0.1240	0.3157	0.1448	0.1935	0.1633	0.1231	0.1715	-3.3029
-4.2213	2.0	21700	0.0247	0.0014	0.0708	0.0181	0.0211	0.0465	0.0189	0.0166	0.0262	0.0199	0.0132	0.0164	0.0208	0.0124	0.0000	0.0001	0.0312	0.0000	0.0002	0.0006	0.0001	0.0006	0.0001	0.0007	0.0001	0.0010	0.0262	0.0023	0.7033	0.0134	0.0710	0.0192	0.0467	0.0201	0.0259	0.0208	0.0158	0.0212	-3.3665
-4.3344	2.9998	32547	0.0057	0.0005	0.0146	0.0038	0.0048	0.0099	0.0041	0.0040	0.0062	0.0047	0.0037	0.0043	0.0055	0.0039	0.0	0.0	0.0	0.0	0.0001	0.0004	0.0001	0.0005	0.0001	0.0006	0.0001	0.0008	0.0060	0.0008	0.5636	0.0030	0.0147	0.0040	0.0102	0.0044	0.0059	0.0048	0.0037	0.0050	-3.3437

Framework versions

• Transformers 4.49.0
• Pytorch 2.7.0
• Datasets 5.0.1
• Tokenizers 0.21.4