https://github.com/defgsus/nn-experiments

playground for neural nets
https://github.com/defgsus/nn-experiments

Last synced: 28 days ago
JSON representation

playground for neural nets

• Host: GitHub
• URL: https://github.com/defgsus/nn-experiments
• Owner: defgsus
• Created: 2023-07-22T00:28:47.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2024-03-17T12:19:58.000Z (8 months ago)
• Last Synced: 2024-10-10T15:17:45.595Z (about 1 month ago)
• Language: Jupyter Notebook
• Size: 11.5 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        

Personal testbed and collection of utilities for torch-based 

neural net experiments.

Folder structure:

- [src/](src/) and [test/](src/) is mostly the library part.

  There is a lot of work on torch Datasets, some of it is

  obsoleted by torch's `DataPipes` library which i do NOT yet use.

- [scripts/](scripts/) contains a couple of different experiments.

  The files are pretty similar in structure and vary in dataset,

  model and training details. Most of it superseded by:

- [experiments/](experiments/) contains new experiments based on

  yaml file descriptions. It's an opinionated wrapper around

  my [Trainer class](src/train/trainer.py) that can also run 

  permutations of parameters and log the results.

- [notebooks-cleaned/](notebooks-cleaned/) contains the cleaned (no output)

  version of all the jupyter notebooks used for experimentation. 

  Some of them are useful, some of them are collections of crap from

  middle-of-the-night trial-and-error sessions.

Experiments from the [experiments/](experiments/) folder are run in project root with:

```shell

python exp.py experiments/the-name.yml run

```

Create a github issue if you are interested in playing with it and

need some documentation.

### :strawberry: Logbooks

[experiments/logs/](experiments/logs/) contains logbooks of some experiments

### CLIPig 2 

After the roaring success of [CLIPig](https://github.com/defgsus/clipig) (ha ha)

i started a version that is configured through a graphical UI instead of yaml files.

![CLIPig screenshot](./src/clipig/screenshot.png)

Run it with 

```shell

python src/clipig

```

### 🤗  (Small) Language Model chat

There's a small tool to put prompts into the 

[huggingface language generators](https://huggingface.co/models?pipeline_tag=text-generation) 

using a browser, which is much nicer than a command line interface. 

```shell

python scripts/chat_lm_browser.py [--model ]

# and visit http://127.0.0.1:8000

```

I'm currently having fun with 

```shell

python scripts/chat_lm_browser.py \

    --model microsoft/phi-1_5

    --device auto

    --bits 8

```

`bits` reduces the precision of the weights

([doc](https://huggingface.co/docs/transformers/llm_tutorial_optimization#1-lower-precision))

which reduces memory requirements and let's you actually put some of the models

on a normal GPU.