|
| 1 | +--- |
| 2 | +layout: model |
| 3 | +title: DistilBERT Sequence Classification - Banking77 (distilbert_sequence_classifier_banking77) |
| 4 | +author: John Snow Labs |
| 5 | +name: distilbert_sequence_classifier_banking77 |
| 6 | +date: 2021-11-21 |
| 7 | +tags: [banking, distilbert, en, english, sequence_classification, open_source] |
| 8 | +task: Text Classification |
| 9 | +language: en |
| 10 | +edition: Spark NLP 3.3.3 |
| 11 | +spark_version: 3.0 |
| 12 | +supported: true |
| 13 | +article_header: |
| 14 | + type: cover |
| 15 | +use_language_switcher: "Python-Scala-Java" |
| 16 | +--- |
| 17 | + |
| 18 | +## Description |
| 19 | + |
| 20 | +Fine-tuned DistilBERT model by using Banking77 dataset. The dataset is composed of online banking queries annotated with their corresponding intents. |
| 21 | + |
| 22 | +BANKING77 dataset provides a very fine-grained set of intents in a banking domain. It comprises 13,083 customer service queries labeled with 77 intents. It focuses on fine-grained single-domain intent detection. |
| 23 | + |
| 24 | +## Predicted Entities |
| 25 | + |
| 26 | +`activate_my_card`, `age_limit`, `apple_pay_or_google_pay`, `atm_support`, `automatic_top_up`, `balance_not_updated_after_bank_transfer`, `balance_not_updated_after_cheque_or_cash_deposit`, `beneficiary_not_allowed`, `cancel_transfer`, `card_about_to_expire`, `card_acceptance`, `card_arrival`, `card_delivery_estimate`, `card_linking`, `card_not_working`, `card_payment_fee_charged`, `card_payment_not_recognised`, `card_payment_wrong_exchange_rate`, `card_swallowed`, `cash_withdrawal_charge`, `cash_withdrawal_not_recognised`, `change_pin`, `compromised_card`, `contactless_not_working`, `country_support`, `declined_card_payment`, `declined_cash_withdrawal`, `declined_transfer`, `direct_debit_payment_not_recognised`, `disposable_card_limits`, `edit_personal_details`, `exchange_charge`, `exchange_rate`, `exchange_via_app`, `extra_charge_on_statement`, `failed_transfer`, `fiat_currency_support`, `get_disposable_virtual_card`, `get_physical_card`, `getting_spare_card`, `getting_virtual_card`, `lost_or_stolen_card`, `lost_or_stolen_phone`, `order_physical_card`, `passcode_forgotten`, `pending_card_payment`, `pending_cash_withdrawal`, `pending_top_up`, `pending_transfer`, `pin_blocked`, `receiving_money`, `Refund_not_showing_up`, `request_refund`, `reverted_card_payment?`, `supported_cards_and_currencies`, `terminate_account`, `top_up_by_bank_transfer_charge`, `top_up_by_card_charge`, `top_up_by_cash_or_cheque`, `top_up_failed`, `top_up_limits`, `top_up_reverted`, `topping_up_by_card`, `transaction_charged_twice`, `transfer_fee_charged`, `transfer_into_account`, `transfer_not_received_by_recipient`, `transfer_timing`, `unable_to_verify_identity`, `verify_my_identity`, `verify_source_of_funds`, `verify_top_up`, `virtual_card_not_working`, `visa_or_mastercard`, `why_verify_identity`, `wrong_amount_of_cash_received`, `wrong_exchange_rate_for_cash_withdrawal` |
| 27 | + |
| 28 | +{:.btn-box} |
| 29 | +<button class="button button-orange" disabled>Live Demo</button> |
| 30 | +<button class="button button-orange" disabled>Open in Colab</button> |
| 31 | +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilbert_sequence_classifier_banking77_en_3.3.3_3.0_1637500452249.zip){:.button.button-orange.button-orange-trans.arr.button-icon} |
| 32 | + |
| 33 | +## How to use |
| 34 | + |
| 35 | + |
| 36 | + |
| 37 | +<div class="tabs-box" markdown="1"> |
| 38 | +{% include programmingLanguageSelectScalaPythonNLU.html %} |
| 39 | +```python |
| 40 | +document_assembler = DocumentAssembler() \ |
| 41 | + .setInputCol('text') \ |
| 42 | + .setOutputCol('document') |
| 43 | + |
| 44 | +tokenizer = Tokenizer() \ |
| 45 | + .setInputCols(['document']) \ |
| 46 | + .setOutputCol('token') |
| 47 | + |
| 48 | +sequenceClassifier = DistilBertForSequenceClassification \ |
| 49 | + .pretrained('distilbert_sequence_classifier_banking77', 'en') \ |
| 50 | + .setInputCols(['token', 'document']) \ |
| 51 | + .setOutputCol('class') \ |
| 52 | + .setMaxSentenceLength(512) |
| 53 | + |
| 54 | +pipeline = Pipeline(stages=[ |
| 55 | + document_assembler, |
| 56 | + tokenizer, |
| 57 | + sequenceClassifier |
| 58 | +]) |
| 59 | + |
| 60 | +example = spark.createDataFrame([['I am still waiting on my card?']]).toDF("text") |
| 61 | +result = pipeline.fit(example).transform(example) |
| 62 | +``` |
| 63 | +```scala |
| 64 | +val document_assembler = DocumentAssembler() |
| 65 | + .setInputCol("text") |
| 66 | + .setOutputCol("document") |
| 67 | + |
| 68 | +val tokenizer = Tokenizer() |
| 69 | + .setInputCols("document") |
| 70 | + .setOutputCol("token") |
| 71 | + |
| 72 | +val tokenClassifier = DistilBertForSequenceClassification.pretrained("distilbert_sequence_classifier_banking77", "en") |
| 73 | + .setInputCols("document", "token") |
| 74 | + .setOutputCol("class") |
| 75 | + .setMaxSentenceLength(512) |
| 76 | + |
| 77 | +val pipeline = new Pipeline().setStages(Array(document_assembler, tokenizer, sequenceClassifier)) |
| 78 | + |
| 79 | +val example = Seq("I am still waiting on my card?").toDS.toDF("text") |
| 80 | + |
| 81 | +val result = pipeline.fit(example).transform(example) |
| 82 | +``` |
| 83 | +</div> |
| 84 | + |
| 85 | +{:.model-param} |
| 86 | +## Model Information |
| 87 | + |
| 88 | +{:.table-model} |
| 89 | +|---|---| |
| 90 | +|Model Name:|distilbert_sequence_classifier_banking77| |
| 91 | +|Compatibility:|Spark NLP 3.3.3+| |
| 92 | +|License:|Open Source| |
| 93 | +|Edition:|Official| |
| 94 | +|Input Labels:|[token, document]| |
| 95 | +|Output Labels:|[class]| |
| 96 | +|Language:|en| |
| 97 | +|Case sensitive:|false| |
| 98 | +|Max sentense length:|512| |
| 99 | + |
| 100 | +## Data Source |
| 101 | + |
| 102 | +[https://huggingface.co/philschmid/DistilBERT-Banking77](https://huggingface.co/philschmid/DistilBERT-Banking77) |
| 103 | + |
| 104 | +[https://huggingface.co/datasets/banking77](https://huggingface.co/datasets/banking77) |
| 105 | + |
| 106 | +## Benchmarking |
| 107 | + |
| 108 | +```bash |
| 109 | +- Loss: 0.2988220155239105 |
| 110 | +- Accuracy: 0.9246753246753247 |
| 111 | +- Macro F1: 0.9246117406953515 |
| 112 | +- Micro F1: 0.9246753246753247 |
| 113 | +- Weighted F1: 0.9246117406953518 |
| 114 | +- Macro Precision: 0.9278163684429038 |
| 115 | +- Micro Precision: 0.9246753246753247 |
| 116 | +- Weighted Precision: 0.927816368442904 |
| 117 | +- Macro Recall: 0.9246753246753248 |
| 118 | +- Micro Recall: 0.9246753246753247 |
| 119 | +- Weighted Recall: 0.9246753246753247 |
| 120 | + |
| 121 | +``` |
0 commit comments