Train AI in 3 lines - The simplest PyTorch alternative
Project description
AL - Alchemy Learning Framework
Train AI in 3 lines. No PhD required.
AL is a lightweight alternative to PyTorch that makes AI accessible to everyone. It packs GPT, transformers, autograd, and GPU support into a single 100KB file with only NumPy as a dependency.
import al
ai = al.quick_train("conversations.json")
print(ai.chat("Hello!"))
Why AL Over PyTorch?
PyTorch is powerful but complex. AL strips away the complexity while keeping the power:
· 3 lines vs 200 lines to train a GPT model · 100KB vs 800MB install size · 5 minutes vs 2 weeks to learn · Runs on phones (Termux/Android) · Same features: autograd, GPU, transformers, AMP
What Can AL Do?
· Train chatbots from JSON conversations · Generate text with GPT-style models · Save/load models in compressed .llk format · Auto-detect GPU (CUDA, MPS) or run on CPU · Mixed precision training for 2x speed · Everything PyTorch does - just simpler
Who Is AL For?
· Students learning AI concepts · Developers prototyping quickly · Hobbyists building weekend projects · Mobile devs running AI on-device · Anyone tired of complex frameworks
Also how do we install it very simple pip install gptperoz==0.1.0
The Philosophy
AI should be fun, not frustrating. AL proves that powerful models don't need complicated code. One file. One dependency. Infinite possibilities.
AL - Because training AI shouldn't require a PhD.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file gptperoz-0.2.2.tar.gz.
File metadata
- Download URL: gptperoz-0.2.2.tar.gz
- Upload date:
- Size: 7.1 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.14.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
4b505857a2143982bac9c965116e799badd5f0b5892c983efffb3780d5d69476
|
|
| MD5 |
b8033cfd0002fc0cc37d214f751640be
|
|
| BLAKE2b-256 |
4b40100c140c32b9b261e7504008499ba526aee7d5556fd01c4f7214942e8dae
|
File details
Details for the file gptperoz-0.2.2-py3-none-any.whl.
File metadata
- Download URL: gptperoz-0.2.2-py3-none-any.whl
- Upload date:
- Size: 7.3 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.14.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
8f5e4a96591f0890e9e990db747a7df7572fa2a68c0abdb5ca2ed49ef90ba35d
|
|
| MD5 |
3d699955e1c18d6a957f99a74328152f
|
|
| BLAKE2b-256 |
0b4ffccdc8c2ae7154ebc01759a0381ce3bb3780361abd815d264b66ffe19110
|