DistilBART_CNN_GNAD_V2
This model is a fine-tuned version of
Einmalumdiewelt/DistilBART_CNN_GNAD_V2
on an unknown dataset.
It achieves the following results on the evaluation set:
-
Loss: 2.7281
-
Rouge1: 27.7253
-
Rouge2: 8.4647
-
Rougel: 18.2059
-
Rougelsum: 23.238
-
Gen Len: 91.6827
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
-
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
-
lr_scheduler_type: linear
-
num_epochs: 10.0
Training results
Framework versions
-
Transformers 4.22.0.dev0
-
Pytorch 1.12.1+cu113
-
Datasets 2.4.0
-
Tokenizers 0.12.1