|
--- |
|
thumbnail: "https://github.com/SauravMaheshkar/CommonLit-Readibility/blob/main/assets/CommonLit%20-%20Big%20Banner.png?raw=true" |
|
tags: |
|
- kaggle |
|
license: "CC0-1.0 License" |
|
datasets: |
|
- Commonlit-Readibility |
|
metrics: |
|
- Perplexity |
|
--- |
|
|
|
![](https://github.com/SauravMaheshkar/CommonLit-Readibility/blob/main/assets/CommonLit%20-%20Big%20Banner.png?raw=true) |
|
|
|
# PreTraining |
|
|
|
| **Architecture** | **Weights** | **PreTraining Loss** | **PreTraining Perplexity** | |
|
|:-----------------------:|:---------------:|:----------------:|:----------------------:| |
|
| roberta-base | [huggingface/hub](https://huggingface.co/SauravMaheshkar/clr-pretrained-roberta-base) | **0.3488** | **3.992** | |
|
| bert-base-uncased | [huggingface/hub](https://huggingface.co/SauravMaheshkar/clr-pretrained-bert-base-uncased) | 0.3909 | 6.122 | |
|
| electra-large | [huggingface/hub](https://huggingface.co/SauravMaheshkar/clr-pretrained-electra-large) | 0.723 | 6.394 | |
|
| albert-base | [huggingface/hub](https://huggingface.co/SauravMaheshkar/clr-pretrained-albert-base) | 0.7343 | 7.76 | |
|
| electra-small | [huggingface/hub](https://huggingface.co/SauravMaheshkar/clr-pretrained-electra-small) | 0.9226 | 11.098 | |
|
| electra-base | [huggingface/hub](https://huggingface.co/SauravMaheshkar/clr-pretrained-electra-base) | 0.9468 | 8.783 | |
|
| distilbert-base-uncased | [huggingface/hub](https://huggingface.co/SauravMaheshkar/clr-pretrained-distilbert-base-uncased) | 1.082 | 7.963 | |
|
|