File size: 4,170 Bytes
e6383f8
777c4ba
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e6383f8
777c4ba
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
711a02a
777c4ba
 
711a02a
777c4ba
 
 
711a02a
777c4ba
 
711a02a
777c4ba
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e150c27
777c4ba
 
e150c27
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
---
language:
- multilingual
- en
- ar
- bg
- de
- el
- es
- fr
- hi
- ru
- sw
- th
- tr
- ur
- vi
- zh
license: cc-by-nc-4.0
datasets:
- xnli
- facebook/anli
pipeline_tag: zero-shot-classification
base_model: facebook/drama-base
model-index:
- name: drama-base-xnli-anli
  results: []
---

# drama-base-xnli-anli

This model is a fine-tuned version of [facebook/drama-base](https://huggingface.co/facebook/drama-base) on the XNLI and ANLI dataset.

## Model description

[DRAMA: Diverse Augmentation from Large Language Models to Smaller Dense Retrievers](https://arxiv.org/abs/2502.18460).
Xueguang Ma, Xi Victoria Lin, Barlas Oguz, Jimmy Lin, Wen-tau Yih, Xilun Chen, arXiv 2025

## How to use the model

### With the zero-shot classification pipeline

The model can be loaded with the `zero-shot-classification` pipeline like so:

```python
from transformers import AutoTokenizer, pipeline
model = "mjwong/drama-base-xnli-anli"
classifier = pipeline("zero-shot-classification",
                      model=model)
```

You can then use this pipeline to classify sequences into any of the class names you specify.

```python
sequence_to_classify = "one day I will see the world"
candidate_labels = ['travel', 'cooking', 'dancing']
classifier(sequence_to_classify, candidate_labels)
```

If more than one candidate label can be correct, pass `multi_class=True` to calculate each class independently:

```python
candidate_labels = ['travel', 'cooking', 'dancing', 'exploration']
classifier(sequence_to_classify, candidate_labels, multi_class=True)
```

### With manual PyTorch

The model can also be applied on NLI tasks like so:

```python
import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification

# device = "cuda:0" or "cpu"
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")

model_name = "mjwong/drama-base-xnli-anli"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

premise = "But I thought you'd sworn off coffee."
hypothesis = "I thought that you vowed to drink more coffee."

input = tokenizer(premise, hypothesis, truncation=True, return_tensors="pt")
output = model(input["input_ids"].to(device))
prediction = torch.softmax(output["logits"][0], -1).tolist()
label_names = ["entailment", "neutral", "contradiction"]
prediction = {name: round(float(pred) * 100, 2) for pred, name in zip(prediction, label_names)}
print(prediction)
```

### Eval results
The model was evaluated using the XNLI test sets on 15 languages: English (en), Arabic (ar), Bulgarian (bg), German (de), Greek (el), Spanish (es), French (fr), Hindi (hi), Russian (ru), Swahili (sw), Thai (th), Turkish (tr), Urdu (ur), Vietnam (vi) and Chinese (zh). The metric used is accuracy.

|Datasets|en|ar|bg|de|el|es|fr|hi|ru|sw|th|tr|ur|vi|zh|
| :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
|[drama-base-xnli-anli](https://huggingface.co/mjwong/drama-base-xnli-anli)|0.788|0.689|0.708|0.715|0.696|0.732|0.737|0.647|0.711|0.636|0.676|0.664|0.588|0.708|0.710|
|[drama-large-xnli-anli](https://huggingface.co/mjwong/drama-large-xnli-anli)|0.799|0.698|0.730|0.721|0.717|0.754|0.754|0.649|0.718|0.652|0.678|0.656|0.594|0.719|0.719|

The model was also evaluated using the dev sets for MultiNLI and test sets for ANLI. The metric used is accuracy.

|Datasets|mnli_dev_m|mnli_dev_mm|anli_test_r1|anli_test_r2|anli_test_r3|
| :---: | :---: | :---: | :---: | :---: | :---: |
|[drama-base-xnli-anli](https://huggingface.co/mjwong/drama-base-xnli-anli)|0.781|0.787|0.500|0.420|0.440|
|[drama-large-xnli-anli](https://huggingface.co/mjwong/drama-large-xnli-anli)|0.794|0.796|0.534|0.446|0.452|

### Training hyperparameters

The following hyperparameters were used during training:

- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1

### Framework versions
- Transformers 4.49.0
- Pytorch 2.6.0+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0