Hugging Face's logo

MoritzLaurer
/
DeBERTa-v3-large-mnli-fever-anli-ling-wanli

Zero-Shot Classification
Transformers
PyTorch
ONNX
Safetensors
English
deberta-v2
text-classification
Eval Results
Inference Endpoints
Model card Files Community
4
DeBERTa-v3-large-mnli-fever-anli-ling-wanli
File size: 10,730 Bytes 
a7ffa6b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
 
---
language: 
- en
tags:
- text-classification
- zero-shot-classification
license: mit
metrics:
- accuracy
datasets:
- multi_nli
- anli
- fever
- lingnli
- alisawuffles/WANLI
#pipeline_tag:
#- text-classification
widget:
- text: "I first thought that I really liked the movie, but upon second thought it was actually disappointing. [SEP] The movie was good."

model-index:  # info: https://github.com/huggingface/hub-docs/blame/main/modelcard.md
- name: DeBERTa-v3-large-mnli-fever-anli-ling-wanli
  results:
  - task:
      type: text-classification             # Required. Example: automatic-speech-recognition
      name: Natural Language Inference             # Optional. Example: Speech Recognition
    dataset:
      type: multi_nli          # Required. Example: common_voice. Use dataset id from https://hf.co/datasets
      name: MultiNLI-matched         # Required. A pretty name for the dataset. Example: Common Voice (French)
      split: validation_matched        # Optional. Example: test
    metrics:
      - type: accuracy         # Required. Example: wer. Use metric id from https://hf.co/metrics
        value: 0,912       # Required. Example: 20.90
        #name:           # Optional. Example: Test WER
        verified: false              # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
  - task:
      type: text-classification             # Required. Example: automatic-speech-recognition
      name: Natural Language Inference             # Optional. Example: Speech Recognition
    dataset:
      type: multi_nli          # Required. Example: common_voice. Use dataset id from https://hf.co/datasets
      name: MultiNLI-mismatched         # Required. A pretty name for the dataset. Example: Common Voice (French)
      split: validation_mismatched        # Optional. Example: test
    metrics:
      - type: accuracy         # Required. Example: wer. Use metric id from https://hf.co/metrics
        value: 0,908       # Required. Example: 20.90
        #name:           # Optional. Example: Test WER
        verified: false              # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
  - task:
      type: text-classification             # Required. Example: automatic-speech-recognition
      name: Natural Language Inference             # Optional. Example: Speech Recognition
    dataset:
      type: anli          # Required. Example: common_voice. Use dataset id from https://hf.co/datasets
      name: ANLI-all         # Required. A pretty name for the dataset. Example: Common Voice (French)
      split: test_r1+test_r2+test_r3        # Optional. Example: test
    metrics:
      - type: accuracy         # Required. Example: wer. Use metric id from https://hf.co/metrics
        value: 0,702       # Required. Example: 20.90
        #name:           # Optional. Example: Test WER
        verified: false              # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
  - task:
      type: text-classification             # Required. Example: automatic-speech-recognition
      name: Natural Language Inference             # Optional. Example: Speech Recognition
    dataset:
      type: anli          # Required. Example: common_voice. Use dataset id from https://hf.co/datasets
      name: ANLI-r3         # Required. A pretty name for the dataset. Example: Common Voice (French)
      split: test_r3        # Optional. Example: test
    metrics:
      - type: accuracy         # Required. Example: wer. Use metric id from https://hf.co/metrics
        value: 0,64       # Required. Example: 20.90
        #name:           # Optional. Example: Test WER
        verified: false              # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
  - task:
      type: text-classification             # Required. Example: automatic-speech-recognition
      name: Natural Language Inference             # Optional. Example: Speech Recognition
    dataset:
      type: alisawuffles/WANLI          # Required. Example: common_voice. Use dataset id from https://hf.co/datasets
      name: WANLI         # Required. A pretty name for the dataset. Example: Common Voice (French)
      split: test        # Optional. Example: test
    metrics:
      - type: accuracy         # Required. Example: wer. Use metric id from https://hf.co/metrics
        value: 0,77       # Required. Example: 20.90
        #name:           # Optional. Example: Test WER
        verified: false              # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
  - task:
      type: text-classification             # Required. Example: automatic-speech-recognition
      name: Natural Language Inference             # Optional. Example: Speech Recognition
    dataset:
      type: lingnli          # Required. Example: common_voice. Use dataset id from https://hf.co/datasets
      name: LingNLI         # Required. A pretty name for the dataset. Example: Common Voice (French)
      split: test        # Optional. Example: test
    metrics:
      - type: accuracy         # Required. Example: wer. Use metric id from https://hf.co/metrics
        value: 0,87       # Required. Example: 20.90
        #name:           # Optional. Example: Test WER
        verified: false              # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).



---

# DeBERTa-v3-large-mnli-fever-anli-ling-wanli
## Model description
This model was fine-tuned on the [MultiNLI](https://huggingface.co/datasets/multi_nli), [Fever-NLI](https://github.com/easonnie/combine-FEVER-NSMN/blob/master/other_resources/nli_fever.md), Adversarial-NLI ([ANLI](https://huggingface.co/datasets/anli)), [LingNLI](https://arxiv.org/pdf/2104.07179.pdf) and [WANLI](https://huggingface.co/datasets/alisawuffles/WANLI) datasets, which comprise 885 242 NLI hypothesis-premise pairs. This model is the best NLI and zero-shot model on the Hugging Face Hub as of 06.06.22. It significantly outperforms all other large models on the [ANLI benchmark](https://github.com/facebookresearch/anli).

The foundation model is [DeBERTa-v3-large from Microsoft](https://huggingface.co/microsoft/deberta-v3-large). Released on 06.12.21, DeBERTa-v3-large is currently the best large-sized foundation model for text classification. It combines several recent innovations compared to classical Masked Language Models like BERT, RoBERTa etc., see the [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-large-mnli-fever-anli-ling-wanli"
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
DeBERTa-v3-large-mnli-fever-anli-ling-wanli was trained on the [MultiNLI](https://huggingface.co/datasets/multi_nli), [Fever-NLI](https://github.com/easonnie/combine-FEVER-NSMN/blob/master/other_resources/nli_fever.md), Adversarial-NLI ([ANLI](https://huggingface.co/datasets/anli)), [LingNLI](https://arxiv.org/pdf/2104.07179.pdf) and [WANLI](https://huggingface.co/datasets/alisawuffles/WANLI) datasets, which comprise 885 242 NLI hypothesis-premise pairs. Note that [SNLI](https://huggingface.co/datasets/snli) was explicitly excluded due to quality issues with the dataset. More data does not necessarily make for better NLI models. 

### Training procedure
DeBERTa-v3-large-mnli-fever-anli-ling-wanli was trained using the Hugging Face trainer with the following hyperparameters. Note that longer training with more epochs hurt performance in my tests (overfitting).


```
training_args = TrainingArguments(
    num_train_epochs=4,              # total number of training epochs
    learning_rate=5e-06,
    per_device_train_batch_size=16,   # batch size per device during training
    gradient_accumulation_steps=2,    # doubles the effective batch_size to 32, while decreasing memory requirements
    per_device_eval_batch_size=64,    # batch size for evaluation
    warmup_ratio=0.06,                # number of warmup steps for learning rate scheduler
    weight_decay=0.01,               # strength of weight decay
    fp16=True                        # mixed precision training
)
```

### Eval results
The model was evaluated using the test sets for MultiNLI, ANLI, LingNLI, WANLI and the dev set for Fever-NLI. The metric used is accuracy.
The model achieves state-of-the-art performance on each dataset. Surprisingly, it outperforms the previous [state-of-the-art on ANLI](https://github.com/facebookresearch/anli) (ALBERT-XXL) by 8,3%. I assume that this is because ANLI was created to fool masked language models like RoBERTa (or ALBERT), while DeBERTa-v3 uses a better pre-training objective (RTD), disentangled attention and I fine-tuned it on higher quality NLI data. 

|Datasets|mnli_test_m|mnli_test_mm|anli_test|anli_test_r3|ling_test|wanli_test|
| :---: | :---: | :---: | :---: | :---: | :---: | :---: |
|Accuracy|0.912|0.908|0.702|0.64|0.87|0.77|
|Speed (text/sec, A100 GPU)|696.0|697.0|488.0|425.0|828.0|980.0|

## Limitations and bias
Please consult the original DeBERTa-v3 paper and literature on different NLI datasets for more information on the training data and potential biases. The model will reproduce statistical patterns in the training data. 

### BibTeX entry and citation info
If you want to cite this model, please cite my [preprint on low-resource text classification](https://osf.io/74b8k/) and the original DeBERTa-v3 paper.

### 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 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.