File size: 4,578 Bytes
29c0ae1
 
 
 
 
 
 
 
346d9f9
bfa84bc
29c0ae1
 
 
 
749bba9
 
 
 
29c0ae1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1a3d74d
29c0ae1
1a3d74d
29c0ae1
36e22a9
 
29c0ae1
 
 
 
 
0e77c94
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
---
language: 
- en
tags:
- text-classification
- zero-shot-classification
metrics:
- accuracy
widget:
- text: "I first thought that I liked the movie, but upon second thought the movie was actually disappointing. [SEP] The movie was good."

---
# DeBERTa-v3-small-mnli-fever-docnli-ling-2c
## Model description
This model was trained on 1.279.665 hypothesis-premise pairs from 8 NLI datasets: [MultiNLI](https://huggingface.co/datasets/multi_nli), [Fever-NLI](https://github.com/easonnie/combine-FEVER-NSMN/blob/master/other_resources/nli_fever.md), [LingNLI](https://arxiv.org/abs/2104.07179) and [DocNLI](https://arxiv.org/pdf/2106.09449.pdf) (which includes [ANLI](https://github.com/facebookresearch/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 create more training data. 

The base model is [DeBERTa-v3-small from Microsoft](https://huggingface.co/microsoft/deberta-v3-small). 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](https://arxiv.org/pdf/2006.03654.pdf) as well as the [DeBERTa-V3 paper](https://arxiv.org/abs/2111.09543). 

## Intended uses & limitations
#### How to use the model
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model_name = "MoritzLaurer/DeBERTa-v3-small-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", "neutral", "contradiction"]
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](https://huggingface.co/datasets/multi_nli), [Fever-NLI](https://github.com/easonnie/combine-FEVER-NSMN/blob/master/other_resources/nli_fever.md), [LingNLI](https://arxiv.org/abs/2104.07179) and [DocNLI](https://arxiv.org/pdf/2106.09449.pdf) (which includes [ANLI](https://github.com/facebookresearch/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
---------|----------|---------|----------|----------
0.927 | 0.921 | 0.892 | 0.684 | 0.673


## Limitations and bias
Please consult the original DeBERTa paper and literature on different NLI datasets for potential biases. 

### BibTeX entry and citation info
If you want to cite this model, please cite the original DeBERTa paper, the respective NLI datasets and include a link to this model on the Hugging Face hub. 

### 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](https://www.linkedin.com/in/moritz-laurer/)

### Debugging and issues
Note that DeBERTa-v3 was released recently 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.