Beta
Explore
Marketplace
Neural Labs
Playground
Wallet
Docs
bert-finetuned-ner – AI Model by Wende | AlphaNeural AI
You can deploy this model and start earning money today!
Wende
/
bert-finetuned-ner
like
0
transformers
pytorch
tensorboard
bert
token-classification
generated_from_trainer
conll2003
apache-2.0
model-index
autotrain_compatible
endpoints_compatible
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
bert-finetuned-ner
This model is a fine-tuned version of
bert-base-cased
on the conll2003 dataset. It achieves the following results on the evaluation set:
Loss: 0.0575
Precision: 0.9322
Recall: 0.9505
F1: 0.9413
Accuracy: 0.9860
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: 3
Training results
Training Loss
Epoch
Step
Validation Loss
Precision
Recall
F1
Accuracy
0.2219
1.0
878
0.0716
0.9076
0.9288
0.9181
0.9808
0.0453
2.0
1756
0.0597
0.9297
0.9477
0.9386
0.9852
0.0239
3.0
2634
0.0575
0.9322
0.9505
0.9413
0.9860
Framework versions
Transformers 4.12.3
Pytorch 1.8.2+cu111
Datasets 1.15.1
Tokenizers 0.10.3
