a0ec55d701fabe0975560ce066dce837

This model is a fine-tuned version of google-bert/bert-base-chinese on the dim/tldr_news dataset. It achieves the following results on the evaluation set:

• Loss: 0.9563
• Data Size: 1.0
• Epoch Runtime: 10.0769
• Accuracy: 0.7109
• F1 Macro: 0.7539
• Rouge1: 0.7109
• Rouge2: 0.0
• Rougel: 0.7109
• Rougelsum: 0.7109

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	1.8089	0	1.1677	0.0987	0.0674	0.0987	0.0	0.0987	0.0987
No log	1	178	1.4655	0.0078	1.6930	0.2777	0.0903	0.2770	0.0	0.2770	0.2777
No log	2	356	1.3955	0.0156	1.6207	0.4261	0.2484	0.4268	0.0	0.4261	0.4261
No log	3	534	1.2277	0.0312	1.8511	0.4815	0.2849	0.4815	0.0	0.4815	0.4815
No log	4	712	1.1102	0.0625	2.1718	0.5227	0.3533	0.5227	0.0	0.5227	0.5220
No log	5	890	1.0471	0.125	2.8251	0.6108	0.4203	0.6122	0.0	0.6108	0.6101
0.0705	6	1068	1.0145	0.25	3.8986	0.5732	0.4208	0.5732	0.0	0.5732	0.5724
0.8559	7	1246	0.9666	0.5	5.8445	0.6342	0.5002	0.6349	0.0	0.6349	0.6342
0.7282	8.0	1424	0.8057	1.0	10.4640	0.6705	0.5170	0.6712	0.0	0.6705	0.6697
0.6809	9.0	1602	0.8030	1.0	10.4259	0.6832	0.6553	0.6832	0.0	0.6832	0.6832
0.5605	10.0	1780	0.7570	1.0	9.9348	0.7124	0.7111	0.7124	0.0	0.7124	0.7124
0.4654	11.0	1958	0.9150	1.0	10.0009	0.6918	0.6938	0.6925	0.0	0.6918	0.6918
0.3287	12.0	2136	0.8594	1.0	10.1143	0.7116	0.7434	0.7116	0.0	0.7116	0.7109
0.2687	13.0	2314	1.0862	1.0	9.9927	0.7031	0.7510	0.7031	0.0	0.7038	0.7031
0.2159	14.0	2492	0.9563	1.0	10.0769	0.7109	0.7539	0.7109	0.0	0.7109	0.7109

Framework versions

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