sgugger Marissa commited on
Commit
ed2e1c8
1 Parent(s): feb8a6d

Add model card (#1)

Browse files

- Add model card (632214ef2f41fb6f6de420449f3f164141892eef)
- Update README.md (0ad01e2712ea65b34d7ff949aaba5e9940646b3c)


Co-authored-by: Marissa Gerchick <Marissa@users.noreply.huggingface.co>

Files changed (1) hide show
  1. README.md +136 -0
README.md ADDED
@@ -0,0 +1,136 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - multilingual
4
+ - en
5
+ - fr
6
+ - es
7
+ - de
8
+ - it
9
+ - pt
10
+ - nl
11
+ - sv
12
+ - pl
13
+ - ru
14
+ - ar
15
+ - tr
16
+ - zh
17
+ - ja
18
+ - ko
19
+ - hi
20
+ - vi
21
+ license: cc-by-nc-4.0
22
+ ---
23
+
24
+ # xlm-mlm-17-1280
25
+
26
+ # Table of Contents
27
+
28
+ 1. [Model Details](#model-details)
29
+ 2. [Uses](#uses)
30
+ 3. [Bias, Risks, and Limitations](#bias-risks-and-limitations)
31
+ 4. [Training](#training)
32
+ 5. [Evaluation](#evaluation)
33
+ 6. [Environmental Impact](#environmental-impact)
34
+ 7. [Technical Specifications](#technical-specifications)
35
+ 8. [Citation](#citation)
36
+ 9. [Model Card Authors](#model-card-authors)
37
+ 10. [How To Get Started With the Model](#how-to-get-started-with-the-model)
38
+
39
+
40
+ # Model Details
41
+
42
+ xlm-mlm-17-1280 is the XLM model, which was proposed in [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau, trained on text in 17 languages. The model is a transformer pretrained using a masked language modeling (MLM) objective.
43
+
44
+ ## Model Description
45
+
46
+ - **Developed by:** See [associated paper](https://arxiv.org/abs/1901.07291) and [GitHub Repo](https://github.com/facebookresearch/XLM)
47
+ - **Model type:** Language model
48
+ - **Language(s) (NLP):** 17 languages, see [GitHub Repo](https://github.com/facebookresearch/XLM#the-17-and-100-languages) for full list.
49
+ - **License:** CC-BY-NC-4.0
50
+ - **Related Models:** [xlm-mlm-17-1280](https://huggingface.co/xlm-mlm-17-1280)
51
+ - **Resources for more information:**
52
+ - [Associated paper](https://arxiv.org/abs/1901.07291)
53
+ - [GitHub Repo](https://github.com/facebookresearch/XLM#the-17-and-100-languages)
54
+ - [Hugging Face Multilingual Models for Inference docs](https://huggingface.co/docs/transformers/v4.20.1/en/multilingual#xlm-with-language-embeddings)
55
+
56
+ # Uses
57
+
58
+ ## Direct Use
59
+
60
+ The model is a language model. The model can be used for masked language modeling.
61
+
62
+ ## Downstream Use
63
+
64
+ To learn more about this task and potential downstream uses, see the Hugging Face [fill mask docs](https://huggingface.co/tasks/fill-mask) and the [Hugging Face Multilingual Models for Inference](https://huggingface.co/docs/transformers/v4.20.1/en/multilingual#xlm-with-language-embeddings) docs. Also see the [associated paper](https://arxiv.org/abs/1901.07291).
65
+
66
+ ## Out-of-Scope Use
67
+
68
+ The model should not be used to intentionally create hostile or alienating environments for people.
69
+
70
+ # Bias, Risks, and Limitations
71
+
72
+ Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)).
73
+
74
+ ## Recommendations
75
+
76
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.
77
+
78
+ # Training
79
+
80
+ This model is the XLM model trained on text in 17 languages. The preprocessing included tokenization and byte-pair-encoding. See the [GitHub repo](https://github.com/facebookresearch/XLM#the-17-and-100-languages) and the [associated paper](https://arxiv.org/pdf/1911.02116.pdf) for further details on the training data and training procedure.
81
+
82
+ [Conneau et al. (2020)](https://arxiv.org/pdf/1911.02116.pdf) report that this model has 16 layers, 1280 hidden states, 16 attention heads, and the dimension of the feed-forward layer is 1520. The vocabulary size is 200k and the total number of parameters is 570M (see Table 7).
83
+
84
+ # Evaluation
85
+
86
+ ## Testing Data, Factors & Metrics
87
+
88
+ The model developers evaluated the model on the XNLI cross-lingual classification task (see the [XNLI data card](https://huggingface.co/datasets/xnli) for more details on XNLI) using the metric of test accuracy. See the [GitHub Repo](https://arxiv.org/pdf/1911.02116.pdf) for further details on the testing data, factors and metrics.
89
+
90
+ ## Results
91
+
92
+ For xlm-mlm-17-1280, the test accuracy on the XNLI cross-lingual classification task in English (en), Spanish (es), German (de), Arabic (ar), and Chinese (zh):
93
+
94
+ |Language| en | es | de | ar | zh |
95
+ |:------:|:--:|:---:|:--:|:--:|:--:|
96
+ | |84.8|79.4 |76.2|71.5|75 |
97
+
98
+ See the [GitHub repo](https://github.com/facebookresearch/XLM#ii-cross-lingual-language-model-pretraining-xlm) for further details.
99
+
100
+ # Environmental Impact
101
+
102
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
103
+
104
+ - **Hardware Type:** More information needed
105
+ - **Hours used:** More information needed
106
+ - **Cloud Provider:** More information needed
107
+ - **Compute Region:** More information needed
108
+ - **Carbon Emitted:** More information needed
109
+
110
+ # Technical Specifications
111
+
112
+ [Conneau et al. (2020)](https://arxiv.org/pdf/1911.02116.pdf) report that this model has 16 layers, 1280 hidden states, 16 attention heads, and the dimension of the feed-forward layer is 1520. The vocabulary size is 200k and the total number of parameters is 570M (see Table 7).
113
+
114
+ # Citation
115
+
116
+ **BibTeX:**
117
+
118
+ ```bibtex
119
+ @article{lample2019cross,
120
+ title={Cross-lingual language model pretraining},
121
+ author={Lample, Guillaume and Conneau, Alexis},
122
+ journal={arXiv preprint arXiv:1901.07291},
123
+ year={2019}
124
+ }
125
+ ```
126
+
127
+ **APA:**
128
+ - Lample, G., & Conneau, A. (2019). Cross-lingual language model pretraining. arXiv preprint arXiv:1901.07291.
129
+
130
+ # Model Card Authors
131
+
132
+ This model card was written by the team at Hugging Face.
133
+
134
+ # How to Get Started with the Model
135
+
136
+ More information needed. See the [ipython notebook](https://github.com/facebookresearch/XLM/blob/main/generate-embeddings.ipynb) in the associated [GitHub repo](https://github.com/facebookresearch/XLM#the-17-and-100-languages) for examples.