Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
legal-roberta-base-filtered-cuad – AI Model by alex-apostolo | AlphaNeural AI
You can deploy this model and start earning money today!
alex-apostolo
/
legal-roberta-base-filtered-cuad
like
0
transformers
pytorch
tensorboard
roberta
question-answering
generated_from_trainer
alex-apostolo/filtered-cuad
apache-2.0
endpoints_compatible
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
legal-roberta-base-filtered-cuad
This model is a fine-tuned version of
saibo/legal-roberta-base
on the cuad dataset. It achieves the following results on the evaluation set:
Loss: 0.0428
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: 2e-05
train_batch_size: 22
eval_batch_size: 22
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 3
Training results
Training Loss
Epoch
Step
Validation Loss
0.0556
1.0
12279
0.0517
0.0406
2.0
24558
0.0425
0.0332
3.0
36837
0.0428
Framework versions
Transformers 4.21.1
Pytorch 1.12.0+cu113
Datasets 2.4.0
Tokenizers 0.12.1
