|
--- |
|
language: en |
|
tags: |
|
- text-classification |
|
- pytorch |
|
- tensorflow |
|
datasets: |
|
- go_emotions |
|
license: mit |
|
--- |
|
|
|
# distilbert-base-uncased-go-emotions-student |
|
|
|
## Model Description |
|
|
|
This model is distilled from the zero-shot classification pipeline on the unlabeled GoEmotions dataset using [this |
|
script](https://github.com/huggingface/transformers/tree/master/examples/research_projects/zero-shot-distillation). |
|
It is the result of the demo notebook |
|
[here](https://colab.research.google.com/drive/1mjBjd0cR8G57ZpsnFCS3ngGyo5nCa9ya?usp=sharing), where more details |
|
about the model can be found. |
|
|
|
- Teacher model: [roberta-large-mnli](https://huggingface.co/roberta-large-mnli) |
|
|
|
## Intended Usage |
|
|
|
The model can be used like any other model trained on GoEmotions, but will likely not perform as well as a model |
|
trained with full supervision. It is primarily intended as a demo of how an expensive NLI-based zero-shot model |
|
can be distilled to a more efficient student. Note that although the GoEmotions dataset allow multiple labels |
|
per instance, the teacher used single-label classification to create psuedo-labels. |
|
|