ELECTRA-small-OWT

This is an unnoficial implementation of an ELECTRA small model, trained on the OpenWebText corpus.

Differences from official ELECTRA models:

• we use a BertForMaskedLM as the generator and BertForTokenClassification as the discriminator
• they use an embedding projection layer, but Bert doesn't have one

Pretraining ttask

(figure from Clark et al. 2020)

ELECTRA uses discriminative LM / replaced-token-detection for pretraining. This involves a generator (a Masked LM model) creating examples for a discriminator to classify as original or replaced for each token.

Usage

from transformers import BertForSequenceClassification, BertTokenizer

tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
electra = BertForSequenceClassification.from_pretrained('shoarora/electra-small-owt')

Code

The pytorch module that implements this task is available here.

Further implementation information here, and here is the script that created this model.

This specific model was trained with the following params:

• batch_size: 512
• training_steps: 5e5
• warmup_steps: 4e4
• learning_rate: 2e-3

Downstream tasks

GLUE Dev results

• Table initialized from ELECTRA github repo

GLUE Test results