da7cdeeed87b40844fd8505728662bd0

This model is a fine-tuned version of openai-community/gpt2-large on the dim/tldr_news dataset. It achieves the following results on the evaluation set:

• Loss: 1.4956
• Data Size: 1.0
• Epoch Runtime: 45.4406
• Accuracy: 0.7798
• F1 Macro: 0.8145
• Rouge1: 0.7798
• Rouge2: 0.0
• Rougel: 0.7798
• Rougelsum: 0.7798

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	1.7365	0	3.7618	0.2429	0.0942	0.2429	0.0	0.2429	0.2429
No log	1	178	1.2177	0.0078	4.7049	0.4822	0.2859	0.4830	0.0	0.4822	0.4830
No log	2	356	0.9219	0.0156	5.3708	0.6513	0.4536	0.6527	0.0	0.6520	0.6520
No log	3	534	0.8017	0.0312	6.7871	0.6832	0.5576	0.6839	0.0	0.6832	0.6832
No log	4	712	0.6968	0.0625	8.8800	0.7422	0.5771	0.7429	0.0	0.7429	0.7429
No log	5	890	0.6977	0.125	12.6272	0.7479	0.7187	0.7479	0.0	0.7486	0.7486
0.0445	6	1068	0.5926	0.25	17.6115	0.7663	0.7351	0.7663	0.0	0.7670	0.7670
0.4515	7	1246	0.6147	0.5	27.5746	0.7706	0.7984	0.7713	0.0	0.7713	0.7706
0.35	8.0	1424	0.5806	1.0	47.0481	0.7869	0.8208	0.7869	0.0	0.7869	0.7869
0.1364	9.0	1602	0.9802	1.0	44.4050	0.7685	0.8079	0.7692	0.0	0.7692	0.7692
0.0652	10.0	1780	1.0321	1.0	44.6383	0.7635	0.8058	0.7642	0.0	0.7628	0.7638
0.0268	11.0	1958	1.4787	1.0	44.3270	0.7635	0.7953	0.7642	0.0	0.7635	0.7635
0.0172	12.0	2136	1.4956	1.0	45.4406	0.7798	0.8145	0.7798	0.0	0.7798	0.7798

Framework versions

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