|
|
--- |
|
|
language: python |
|
|
tags: vae |
|
|
license: apache-2.0 |
|
|
datasets: Fraser/python-lines |
|
|
--- |
|
|
|
|
|
# T5-VAE-Python (flax) |
|
|
|
|
|
A Transformer-VAE made using flax. |
|
|
|
|
|
It has been trained to interpolate on lines of Python code form the [python-lines dataset](https://huggingface.co/datasets/Fraser/python-lines). |
|
|
|
|
|
Done as part of Huggingface community training ([see forum post](https://discuss.huggingface.co/t/train-a-vae-to-interpolate-on-english-sentences/7548)). |
|
|
|
|
|
Builds on T5, using an autoencoder to convert it into an MMD-VAE. |
|
|
|
|
|
## How to use from the 🤗/transformers library |
|
|
|
|
|
Clone the model repo. |
|
|
```bash |
|
|
git clone https://github.com/Fraser-Greenlee/t5-vae-flax.git |
|
|
``` |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer |
|
|
from t5_vae_flax.src.t5_vae import FlaxT5VaeForAutoencoding |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("t5-base") |
|
|
|
|
|
model = T5VaeForAutoencoding.from_pretrained("flax-community/t5-vae-python") |
|
|
``` |
|
|
|
|
|
## Setup |
|
|
|
|
|
Run `setup_tpu_vm_venv.sh` to setup a virtual enviroment on a TPU VM for training. |
|
|
|
