metadata

language:
  - en
license: apache-2.0
library_name: transformers
tags:
  - generated_from_keras_callback
datasets:
  - Babelscape/multinerd
metrics:
  - seqeval
base_model: distilbert-base-uncased
pipeline_tag: token-classification
widget:
  - text: >-
      After months of meticulous review and analysis, I am proud to present a
      study that explores the deep connections between Epstein-Barr virus (EBV),
      Long COVID and Myalgic Encephalomyelitis.
    example_title: Example 1
  - text: Is it dangerous for a tarantula to live in a paludarium?
    example_title: Example 2
  - text: Billionaire Charlie Munger, Warren Buffet's right hand man, dies at 99.
    example_title: Example 3
model-index:
  - name: i-be-snek/distilbert-base-uncased-finetuned-ner-exp_B
    results:
      - task:
          type: token-classification
          name: ner
        dataset:
          name: Babelscape/multinerd (modified version)
          type: Babelscape/multinerd
          split: test
        metrics:
          - type: seqeval
            value: 0.9362959157462112
            name: precision
          - type: seqeval
            value: 0.9524846478811898
            name: recall
          - type: seqeval
            value: 0.9443209050281742
            name: f1
          - type: seqeval
            value: 0.9913435631438657
            name: accuracy

i-be-snek/distilbert-base-uncased-finetuned-ner-exp_B

This model is a fine-tuned version of distilbert-base-uncased on the English subset of the NER Babelscape/multinerd dataset.

It achieves the following results on the validation set:

Train Loss: 0.0084
Validation Loss: 0.0587
Train Precision: 0.9185
Train Recall: 0.9240
Train F1: 0.9213
Train Accuracy: 0.9857
Epoch: 2

Model description

distilbert-base-uncased-finetuned-ner-exp_B is a Named Entity Recognition model finetuned on distilbert-base-uncased. This model is uncased, so it makes no distinction between "sarah" and "Sarah". The dataset it was fine-tuned on was modified. Only five entities were considered: Person (PER), Animal (ANIM), Organization (ORG), Location (LOC), and Disease (DIS). The dataset was modified further so that all other named entities not included in this list were swapped with the '0' label ID. Tokens IDs were also re-indexed.

Training and evaluation data

This model has been evaluated on the English subset of the test set of Babelscape/multinerd with modifications where all tags other than (0) Person (PER), Animal (ANIM), Organization (ORG), Location (LOC), and Disease (DIS) were replaced with the 'O' tag. The label indices were also reset and the dataset was transformed accordingly. You can preprocess the dataset in the same way with any custom set of tags using the script in this GitHub repository

Evaluation results

metric	value
precision	0.936296
recall	0.952485
f1	0.944321
accuracy	0.991344

metric/tag	ANIM	DIS	LOC	ORG	PER
precision	0.674603	0.695304	0.966669	0.954712	0.989048
recall	0.794888	0.799736	0.967232	0.942883	0.994872
f1	0.729823	0.743873	0.966951	0.948761	0.991952
number	3208	1518	24048	6618	10530

Training procedure

All scripts for training can be found in this GitHub repository. The model had early stopped watching its val_loss.

Training hyperparameters

The following hyperparameters were used during training:

optimizer:

  {
    "name": "AdamWeightDecay",
    "learning_rate": 2e-05,
    "decay": 0.0,
    "beta_1": 0.9,
    "beta_2": 0.999,
    "epsilon": 1e-07,
    "amsgrad": False,
    "weight_decay_rate": 0.0,
}

training_precision: float32

Training results

Train Loss	Validation Loss	Train Precision	Train Recall	Train F1	Train Accuracy	Epoch
0.0426	0.0401	0.9159	0.9284	0.9221	0.9860	0
0.0163	0.0451	0.9275	0.9235	0.9255	0.9865	1
0.0084	0.0587	0.9185	0.9240	0.9213	0.9857	2

Epoch 0

Named Entity	precision	recall	f1
ANIM	0.661526	0.741578	0.699269
DIS	0.722194	0.763900	0.742462
LOC	0.965829	0.974215	0.970004
ORG	0.949038	0.906056	0.927049
PER	0.988075	0.989184	0.988629

Epoch 1

Named Entity	precision	recall	f1
ANIM	0.704151	0.646720	0.674214
DIS	0.750533	0.756379	0.753445
LOC	0.969905	0.973037	0.971468
ORG	0.930323	0.932971	0.931645
PER	0.991814	0.989082	0.990446

Epoch 2

Named Entity	precision	recall	f1
ANIM	0.689281	0.675532	0.682337
DIS	0.703917	0.786731	0.743024
LOC	0.975097	0.960122	0.967552
ORG	0.928553	0.931677	0.930112
PER	0.992620	0.988163	0.990387

Framework versions

Transformers 4.35.2
TensorFlow 2.14.0
Datasets 2.15.0
Tokenizers 0.15.0