https://github.com/revsic/tf-neural-process
Tensorflow implementation of Neural Process family
https://github.com/revsic/tf-neural-process
Last synced: 14 days ago
JSON representation
Tensorflow implementation of Neural Process family
- Host: GitHub
- URL: https://github.com/revsic/tf-neural-process
- Owner: revsic
- License: mit
- Created: 2019-04-27T19:15:54.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2019-05-28T18:17:54.000Z (almost 6 years ago)
- Last Synced: 2025-05-05T17:33:13.236Z (14 days ago)
- Language: Jupyter Notebook
- Homepage:
- Size: 2.33 MB
- Stars: 6
- Watchers: 4
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# tf-neural-process
Tensorflow implementation of Neural Process family
- Full code and experiment is based on deepmind repository [[GIT](https://github.com/deepmind/neural-processes), [LICENSE](./3RD-PARTY.md)]
- This repository aims to write more generic code and split model from experiment.
- Paper:
1. Conditional Neural Process [[arxiv](https://arxiv.org/abs/1807.01613)]
2. Neural Process [[arxiv](https://arxiv.org/abs/1807.01622)]
3. Attentive Neural Process [[arxiv](https://arxiv.org/abs/1901.05761)]
## Sample
Jupyter notebook sample is [here](./NeuralProcess.ipynb).
### 1. Conditional Neural Process
Model definition [[GIT](./neural_process/cnp.py)]
```python
encoder_output_sizes = [128, 128, 128, 128]
decoder_output_sizes = [128, 128, 1]
model = neural_process.ConditionalNP(encoder_output_sizes, decoder_output_sizes)
```
Sample image
### 2. Neural Process
Model definition [[GIT](./neural_process/np.py)]
```python
z_output_sizes = [128, 128, 128, 128]
enc_output_sizes = [128, 128, 128, 128]
dec_output_sizes = [128, 128, 1]
model = neural_process.NeuralProcess(z_output_sizes, enc_output_sizes, dec_output_sizes)
```
Sample image
### 3. Attentive Neural Process
Model definition [[GIT](./neural_process/anp.py)]
```python
z_output_sizes = [128, 128, 128, 128]
enc_output_sizes = [128, 128, 128, 128]
cross_output_sizes = [128, 128, 128, 128]
dec_output_sizes = [128, 128, 1]
self_attention = neural_process.Attention(attention_type='multihead', proj=[128, 128])
cross_attention = neural_process.Attention(attention_type='multihead', proj=[128, 128])
model = neural_process.AttentiveNP(z_output_sizes,
enc_output_sizes,
cross_output_sizes,
dec_output_sizes,
self_attention,
cross_attention)
```
Sample image
