Model description
This model was trained on 1.279.665 hypothesis-premise pairs from 8 NLI datasets: MultiNLI, Fever-NLI, LingNLI and DocNLI (which includes ANLI, QNLI, DUC, CNN/DailyMail, Curation).
It is the only model in the model hub trained on 8 NLI datasets, including DocNLI with very long texts to learn long range reasoning. Note that the model was trained on binary NLI to predict either "entailment" or "not-entailment". The DocNLI merges the classes "neural" and "contradiction" into "not-entailment" to enable the inclusion of the DocNLI dataset.
The base model is DeBERTa-v3-base from Microsoft. The v3 variant of DeBERTa substantially outperforms previous versions of the model by including a different pre-training objective, see annex 11 of the original DeBERTa paper as well as the DeBERTa-V3 paper.
For highest performance (but less speed), I recommend using https://huggingface.co/MoritzLaurer/DeBERTa-v3-large-mnli-fever-anli-ling-wanli.
How to use the model
Simple zero-shot classification pipeline
from transformers import pipeline
classifier = pipeline("zero-shot-classification", model="MoritzLaurer/DeBERTa-v3-base-mnli-fever-docnli-ling-2c")
sequence_to_classify = "Angela Merkel is a politician in Germany and leader of the CDU"
candidate_labels = ["politics", "economy", "entertainment", "environment"]
output = classifier(sequence_to_classify, candidate_labels, multi_label=False)
print(output)
NLI use-case
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
model_name = "MoritzLaurer/DeBERTa-v3-base-mnli-fever-docnli-ling-2c"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)
premise = "I first thought that I liked the movie, but upon second thought it was actually disappointing."
hypothesis = "The movie was good."
input = tokenizer(premise, hypothesis, truncation=True, return_tensors="pt")
output = model(input["input_ids"].to(device)) # device = "cuda:0" or "cpu"
prediction = torch.softmax(output["logits"][0], -1).tolist()
label_names = ["entailment", "not_entailment"]
prediction = {name: round(float(pred) * 100, 1) for pred, name in zip(prediction, label_names)}
print(prediction)
Training data
This model was trained on 1.279.665 hypothesis-premise pairs from 8 NLI datasets: MultiNLI, Fever-NLI, LingNLI and DocNLI (which includes ANLI, QNLI, DUC, CNN/DailyMail, Curation).
Training procedure
DeBERTa-v3-small-mnli-fever-docnli-ling-2c was trained using the Hugging Face trainer with the following hyperparameters.
training_args = TrainingArguments(
num_train_epochs=3, # total number of training epochs
learning_rate=2e-05,
per_device_train_batch_size=32, # batch size per device during training
per_device_eval_batch_size=32, # batch size for evaluation
warmup_ratio=0.1, # number of warmup steps for learning rate scheduler
weight_decay=0.06, # strength of weight decay
fp16=True # mixed precision training
)
Eval results
The model was evaluated using the binary test sets for MultiNLI and ANLI and the binary dev set for Fever-NLI (two classes instead of three). The metric used is accuracy.
| mnli-m-2c |
mnli-mm-2c |
fever-nli-2c |
anli-all-2c |
anli-r3-2c |
lingnli-2c |
| 0.935 |
0.933 |
0.897 |
0.710 |
0.678 |
0.895 |
Limitations and bias
Please consult the original DeBERTa paper and literature on different NLI datasets for potential biases.
Citation
If you use this model, please cite: Laurer, Moritz, Wouter van Atteveldt, Andreu Salleras Casas, and Kasper Welbers. 2022. ‘Less Annotating, More Classifying – Addressing the Data Scarcity Issue of Supervised Machine Learning with Deep Transfer Learning and BERT - NLI’. Preprint, June. Open Science Framework. https://osf.io/74b8k.
Ideas for cooperation or questions?
If you have questions or ideas for cooperation, contact me at m{dot}laurer{at}vu{dot}nl or LinkedIn
Debugging and issues
Note that DeBERTa-v3 was released on 06.12.21 and older versions of HF Transformers seem to have issues running the model (e.g. resulting in an issue with the tokenizer). Using Transformers>=4.13 might solve some issues.
