880c68f5d67eb9c8f26b228efcb2b361

This model is a fine-tuned version of meta-llama/Llama-3.2-3B on the dim/tldr_news dataset. It achieves the following results on the evaluation set:

• Loss: 4.8009
• Data Size: 1.0
• Epoch Runtime: 71.0921
• Accuracy: 0.7436
• F1 Macro: 0.7847
• Rouge1: 0.7436
• Rouge2: 0.0
• Rougel: 0.7436
• Rougelsum: 0.7436

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	8.9187	0	4.5927	0.1790	0.1325	0.1783	0.0	0.1790	0.1790
No log	1	178	20.9395	0.0078	6.2743	0.2940	0.1205	0.2933	0.0	0.2933	0.2940
No log	2	356	5.1233	0.0156	7.3387	0.6207	0.4369	0.6222	0.0	0.6218	0.6214
No log	3	534	5.4966	0.0312	12.7083	0.5625	0.4144	0.5618	0.0	0.5629	0.5618
No log	4	712	4.5537	0.0625	16.2372	0.7195	0.5636	0.7205	0.0	0.7202	0.7195
No log	5	890	3.5543	0.125	23.4999	0.6882	0.6144	0.6889	0.0	0.6889	0.6875
0.3277	6	1068	2.9529	0.25	33.2491	0.7138	0.6940	0.7145	0.0	0.7138	0.7145
2.5427	7	1246	2.6212	0.5	48.3250	0.7422	0.7493	0.7429	0.0	0.7429	0.7422
2.0437	8.0	1424	2.6727	1.0	77.7889	0.7472	0.7305	0.7475	0.0	0.7472	0.7472
1.2815	9.0	1602	3.3272	1.0	65.5149	0.7372	0.7377	0.7379	0.0	0.7372	0.7372
0.604	10.0	1780	4.6799	1.0	64.4114	0.7315	0.7577	0.7322	0.0	0.7315	0.7315
0.7612	11.0	1958	4.8009	1.0	71.0921	0.7436	0.7847	0.7436	0.0	0.7436	0.7436

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1