bert-small-finetuned-cuad-longer

muhtasham/bert-small-finetuned-cuad-longer

bert-small-finetuned-cuad-longer is a question answering model from muhtasham. Use it when the input is a question plus a passage. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

15% of all-time downloads

All-time downloads

660

Public

Parameters

28.5M

10.4 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin114 MB · 50%

Parameter types

How the weights are stored.

F3228.5M · 100%

Try a prompt

Base models

google/bert_uncased_L-4_H-512_A-8

Datasets

cuad

Model card

From the Hugging Face model README

bert-small-finetuned-cuad-longer

This model is a fine-tuned version of google/bert_uncased_L-4_H-512_A-8 on the cuad dataset. It achieves the following results on the evaluation set:

Loss: 0.0617

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: 16
eval_batch_size: 16
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
0.0601	1.0	8702	0.0404
0.0551	2.0	17404	0.0394
0.0428	3.0	26106	0.0481
0.0375	4.0	34808	0.0425
0.0403	5.0	43510	0.0551
0.0246	6.0	52212	0.0588
0.0284	7.0	60914	0.0557
0.0303	8.0	69616	0.0543
0.0239	9.0	78318	0.0634
0.0207	10.0	87020	0.0617

Framework versions

Transformers 4.21.1
Pytorch 1.12.1+cu113
Datasets 2.4.0
Tokenizers 0.12.1

Task: Question Answering
Library: transformers
Type: bert
License: apache-2.0
Created: Aug 13, 2022
Updated: Sep 15, 2023