MoritzLaurer HF staff commited on
Commit
35cdaef
1 Parent(s): 4954f85

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +7 -3
README.md CHANGED
@@ -1,6 +1,7 @@
1
  ---
2
  language:
3
  - en
 
4
  tags:
5
  - text-classification
6
  - zero-shot-classification
@@ -18,11 +19,14 @@ pipeline_tag: zero-shot-classification
18
  This model was trained on the MultiNLI, Fever-NLI and Adversarial-NLI (ANLI) datasets, which comprise 763 913 NLI hypothesis-premise pairs. This base model outperforms almost all large models on the [ANLI benchmark](https://github.com/facebookresearch/anli).
19
  The base model is [DeBERTa-v3-base from Microsoft](https://huggingface.co/microsoft/deberta-v3-base). 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).
20
 
 
 
21
  ## Intended uses & limitations
22
  #### How to use the model
23
  ```python
24
  from transformers import AutoTokenizer, AutoModelForSequenceClassification
25
  import torch
 
26
 
27
  model_name = "MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli"
28
  tokenizer = AutoTokenizer.from_pretrained(model_name)
@@ -64,11 +68,11 @@ mnli-m | mnli-mm | fever-nli | anli-all | anli-r3
64
  ## Limitations and bias
65
  Please consult the original DeBERTa paper and literature on different NLI datasets for potential biases.
66
 
67
- ### BibTeX entry and citation info
68
- 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.
69
 
70
  ### Ideas for cooperation or questions?
71
  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/)
72
 
73
  ### Debugging and issues
74
- 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.
 
1
  ---
2
  language:
3
  - en
4
+ license: mit
5
  tags:
6
  - text-classification
7
  - zero-shot-classification
 
19
  This model was trained on the MultiNLI, Fever-NLI and Adversarial-NLI (ANLI) datasets, which comprise 763 913 NLI hypothesis-premise pairs. This base model outperforms almost all large models on the [ANLI benchmark](https://github.com/facebookresearch/anli).
20
  The base model is [DeBERTa-v3-base from Microsoft](https://huggingface.co/microsoft/deberta-v3-base). 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).
21
 
22
+ For highest performance (but less speed), I recommend using https://huggingface.co/MoritzLaurer/DeBERTa-v3-large-mnli-fever-anli-ling-wanli.
23
+
24
  ## Intended uses & limitations
25
  #### How to use the model
26
  ```python
27
  from transformers import AutoTokenizer, AutoModelForSequenceClassification
28
  import torch
29
+ device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
30
 
31
  model_name = "MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli"
32
  tokenizer = AutoTokenizer.from_pretrained(model_name)
 
68
  ## Limitations and bias
69
  Please consult the original DeBERTa paper and literature on different NLI datasets for potential biases.
70
 
71
+ ## Citation
72
+ 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.
73
 
74
  ### Ideas for cooperation or questions?
75
  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/)
76
 
77
  ### Debugging and issues
78
+ 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.