19091b79c23aa18d1e86cb8d3cce32cd

This model is a fine-tuned version of meta-llama/Llama-3.2-1B-Instruct on the dim/tldr_news dataset. It achieves the following results on the evaluation set:

• Loss: 6.6490
• Data Size: 1.0
• Epoch Runtime: 29.0967
• Accuracy: 0.7351
• F1 Macro: 0.7785
• Rouge1: 0.7358
• Rouge2: 0.0
• Rougel: 0.7351
• Rougelsum: 0.7351

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	11.0898	0	2.7847	0.1648	0.1248	0.1641	0.0	0.1648	0.1641
No log	1	178	22.2126	0.0078	2.7055	0.2571	0.1008	0.2571	0.0	0.2578	0.2564
No log	2	356	5.7174	0.0156	4.1475	0.4240	0.2367	0.4240	0.0	0.4240	0.4240
No log	3	534	5.1037	0.0312	6.3905	0.5646	0.4126	0.5639	0.0	0.5646	0.5646
No log	4	712	3.9046	0.0625	7.5815	0.7088	0.5644	0.7095	0.0	0.7095	0.7088
No log	5	890	3.2202	0.125	9.1255	0.7053	0.6141	0.7067	0.0	0.7053	0.7060
0.3024	6	1068	2.7916	0.25	12.1653	0.7280	0.6380	0.7294	0.0	0.7280	0.7287
2.5771	7	1246	2.9160	0.5	19.0042	0.7280	0.7155	0.7287	0.0	0.7290	0.7280
1.9239	8.0	1424	2.7414	1.0	32.4775	0.75	0.7325	0.7514	0.0	0.7507	0.75
0.9656	9.0	1602	3.6120	1.0	32.0793	0.7528	0.7548	0.7539	0.0	0.7525	0.7536
0.6819	10.0	1780	5.4613	1.0	34.0914	0.7550	0.7944	0.7557	0.0	0.7557	0.7550
0.5557	11.0	1958	6.7016	1.0	29.0314	0.7401	0.7616	0.7401	0.0	0.7401	0.7401
0.5357	12.0	2136	6.6490	1.0	29.0967	0.7351	0.7785	0.7358	0.0	0.7351	0.7351

Framework versions

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