roberta2roberta-roberta-large-cnn-dailymail-seed42

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 17.5411
• Rouge1: 0.0370
• Rouge2: 0.0003
• Rougel: 0.0314
• Rougelsum: 0.0347

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: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
6.6238	0.2229	2000	6.5089	0.1704	0.0101	0.1343	0.1651
6.4347	0.4458	4000	16.0748	0.0139	0.0001	0.0125	0.0133
6.2871	0.6687	6000	17.4860	0.0226	0.0001	0.0198	0.0213
6.2286	0.8916	8000	16.4178	0.0515	0.0003	0.0407	0.0466
6.0924	1.1145	10000	15.7300	0.0262	0.0001	0.0227	0.0246
6.0521	1.3374	12000	16.8723	0.0277	0.0002	0.0240	0.0261
6.0173	1.5603	14000	15.3008	0.0374	0.0002	0.0313	0.0348
5.9986	1.7832	16000	16.4679	0.0357	0.0002	0.0302	0.0334
5.963	2.0061	18000	17.3233	0.0332	0.0002	0.0284	0.0312
5.8459	2.2290	20000	16.0664	0.0372	0.0002	0.0314	0.0347
5.8192	2.4519	22000	17.1699	0.0350	0.0002	0.0297	0.0327
5.7961	2.6748	24000	17.4243	0.0336	0.0002	0.0287	0.0314
5.791	2.8977	26000	17.5411	0.0370	0.0003	0.0314	0.0347

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1