2ee1b0d40e90404ca523429e313e2d51

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

• Loss: 1.1165
• Data Size: 1.0
• Epoch Runtime: 10.5807
• Accuracy: 0.7592
• F1 Macro: 0.7925
• Rouge1: 0.7592
• Rouge2: 0.0
• Rougel: 0.7585
• Rougelsum: 0.7592

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	1.5907	0	1.2155	0.2393	0.1190	0.2393	0.0	0.2401	0.2401
No log	1	178	1.5068	0.0078	2.1815	0.3423	0.1925	0.3430	0.0	0.3438	0.3423
No log	2	356	1.3680	0.0156	1.7671	0.3509	0.2085	0.3516	0.0	0.3509	0.3509
No log	3	534	1.0618	0.0312	2.1966	0.6030	0.4541	0.6030	0.0	0.6030	0.6037
No log	4	712	0.8576	0.0625	2.5076	0.6882	0.5286	0.6896	0.0	0.6889	0.6882
No log	5	890	0.7551	0.125	3.1324	0.7081	0.5424	0.7088	0.0	0.7088	0.7077
0.0575	6	1068	0.7392	0.25	4.3357	0.7145	0.5685	0.7152	0.0	0.7152	0.7145
0.6035	7	1246	0.6601	0.5	6.6367	0.7507	0.7441	0.7511	0.0	0.7507	0.7507
0.5065	8.0	1424	0.5901	1.0	11.5137	0.7592	0.7553	0.7599	0.0	0.7599	0.7592
0.3592	9.0	1602	0.7052	1.0	11.3558	0.7536	0.7481	0.7536	0.0	0.7536	0.7543
0.2311	10.0	1780	0.8173	1.0	11.0890	0.7578	0.7839	0.7585	0.0	0.7578	0.7585
0.1485	11.0	1958	0.9321	1.0	10.9751	0.7521	0.7854	0.7521	0.0	0.7521	0.7528
0.164	12.0	2136	1.1165	1.0	10.5807	0.7592	0.7925	0.7592	0.0	0.7585	0.7592

Framework versions

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