Metadata-Version: 2.4
Name: bert-cli
Version: 1.2.0
Summary: Bert — A CLI Framework dev tool for decoder transformer models.
Home-page: https://github.com/mnisperuza/bert-cli
Author: Matias Nisperuza
Author-email: Matias Nisperuza <mnisperuza1102@gmail.com>
License: Apache 2.0
Project-URL: Homepage, https://mnisperuza.github.io/bert-cli/
Project-URL: Documentation, https://github.com/mnisperuza/bert-cli#readme
Project-URL: Repository, https://github.com/mnisperuza/bert-cli
Project-URL: Issues, https://github.com/mnisperuza/bert-cli
Keywords: ai,assistant,llm,local,cli,qwen
Classifier: Development Status :: 5 - Production/Stable
Classifier: Environment :: Console
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: End Users/Desktop
Classifier: Operating System :: OS Independent
Classifier: Operating System :: Microsoft :: Windows
Classifier: Operating System :: POSIX :: Linux
Classifier: Operating System :: MacOS
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: torch>=2.1.0
Requires-Dist: transformers>=4.42.0
Requires-Dist: accelerate>=0.27.0
Requires-Dist: huggingface-hub>=0.20.0
Requires-Dist: hf_xet>=1.1.0
Requires-Dist: safetensors>=0.4.0
Requires-Dist: sentencepiece>=0.1.99
Requires-Dist: einops>=0.7.0
Requires-Dist: numpy<2.0.0,>=1.24.0
Requires-Dist: bitsandbytes>=0.43.0
Provides-Extra: linux
Requires-Dist: bitsandbytes>=0.43.0; extra == "linux"
Provides-Extra: perf
Requires-Dist: xformers>=0.0.25; extra == "perf"
Provides-Extra: full
Requires-Dist: bitsandbytes>=0.43.0; extra == "full"
Requires-Dist: xformers>=0.0.25; extra == "full"
Dynamic: author
Dynamic: home-page
Dynamic: license-file
Dynamic: requires-python

## BERT CLI

**A Educacional Project by Matias Nisperuza**

#### **Give the repository a star if you find it cool or useful!🌟**

![BERT CLI](https://img.shields.io/badge/Bert-CLI-598556?style=for-the-badge)
![Version](https://img.shields.io/badge/version-1.0.0-96C3A3?style=for-the-badge)
![Python](https://img.shields.io/badge/Python-3.8+-3776AB?style=for-the-badge&logo=python&logoColor=white)
![HuggingFace](https://img.shields.io/badge/HuggingFace-%23FFD21E.svg?style=for-the-badge&logo=huggingface&logoColor=white)
![PyTorch](https://img.shields.io/badge/PyTorch-%23EE4C2C.svg?style=for-the-badge&logo=PyTorch&logoColor=white)
![Platform](https://img.shields.io/badge/platform-Windows%20%7C%20Linux%20%7C%20macOS-lightgrey)
![License](https://img.shields.io/badge/License-Apache2.0-blue)

---

## Updates: 
Bert its now at 1.2.0 , Featuring:
- Fixed paste support
- Terminal color themes commands: -color light / -color dark (***they do work***)
- Multiline paste mode: /*paste


Visit berts Official GitHub Page here: [Bert CLI official Page](https://mnisperuza.github.io/bert-cli/)


## ⚖️ Legal & Licensing
This is **Software Licensed Under Apache 2.0**.  
- For licensing terms, see [LICENSE](./LICENSE).

---

## Overview

### About Bert:
![Bert CLI](https://img.shields.io/badge/Bert-598556?style=for-the-badge&logo=data:image/png;base64,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&logoColor=white)

Bert CLI is a small educational dev tool project I’m building and maintaining. Working on the CLI, website, and server has been a fun way for me to learn, break things, and improve my skills while building something cool.

**Bert CLI has nothing to do with BERT (Bidirectional Encoder Representations from Transformers)**, Bert CLI Is just a Framework that host decoder models such as Qwen 2.5/3 models from Alibaba or LFM2 models from LiquidAI.

If you spot any issues, have ideas, or just want to give feedback, open a issue on GitHub, if you want to collaborate, send me an email — I’m always open to it.



<p align="center">
  <img src="assets/preview.png" alt="Bert CLI Preview" width="500">
</p>


---
## Quick Start


### PyPI

```bash
# Directly from PyPI Package:

pip install bert-cli

```

---

## Usage

### Start Bert
```bash
bert
```
#### Claim Your weekly token to start using Bert CLI

Go to [Bert CLI official Page](https://mnisperuza.github.io/bert-cli/).

Claim yout token -

Then run bert and use:
```
/*token YOUR-TOKEN-HERE
```
### Why Tokens Keys?

Token keys may seem unreliable, I know, But I have seen that, to ensure proper chat inside the structure I built with the memory system, a Limit of 20,000 Tokens seems fine, Its Free and Totally open.

I do store your data, but I am not going to do anything with it, I only store it to track and assign a ID to the Tokens, If you have any complain, doubt, Feedback or issue, feel free to email me: mnisperuza1102@gmail.com

>For more info visit:  [Bert CLI official Page](https://amphydia.github.io/bert-cli/)

---

## Models

| Model | Base | VRAM | Features |
|-------|------|------|----------|
| **Bert Nano** | LiquidAI/LFM2-700M | ~2GB | Ultra-fast |
| **Bert Mini** | LiquidAI/LFM2-1.2B | ~4GB | Balanced |
| **Bert Main** | Qwen/Qwen3-1.7B | ~5GB | Thinking 🧠 |
| **Bert Max** | LiquidAI/LFM2-2.6B | ~8GB | Reasoning |
| **Bert Coder** | Qwen/Qwen2.5-Coder-1.5B-Instruct | ~4GB | Code |
| **Bert Max-Coder** | Qwen/Qwen2.5-Coder-3B-Instruct | ~8GB | Heavy Code |

### Command Line Options
```bash
bert --ver      # Show version
bert --info     # Show info
bert --del      # Remove Bert data (~/.bert)
bert --help     # Show help
```

### In-Session Commands

**Switch Models:**
```
bert nano       # Fastest (0.7B)
bert mini       # Balanced (1.2B)
bert main       # Flagship (1.7B)
bert max        # Most capable (2.6B)
bert coder      # Code-optimized (1.5B)
bert maxcoder   # The best for Code (3B)
```

**Change Quantization:**
```
bert int4       # Balanced ⭐ 
bert int8       # High quality 
bert fp16       # Best quality (all platforms)
bert fp32       # Full precision / CPU
```

**Other Commands:**
```
/*help          # Show all commands
/*status        # Show current status
/*clear         # Clear screen
/*exit          # Exit Bert
-color light    # change to Bone white background
-color dark     # change to pitch black
/*paste         # allows multiline paste
```

---

## System Requirements

| Component | Minimum | Recommended |
|-----------|---------|-------------|
| RAM | 8GB | 16GB+ |
| VRAM | 3GB | 6GB+ |
| Python | 3.8 | 3.10+ |
| Storage | 30GB | 40GB |

### Quantization Support

| Platform | INT4/INT8 | FP16 | FP32 |
|----------|-----------|------|------|
| Linux | ✅ | ✅ | ✅ |
| Windows | ✅* | ✅ | ✅ |
| macOS | ❌ | ✅ | ✅ |



---

## Uninstall

Remove Bert data:
```bash
bert --del
```

---

## Support

- **GitHub Issues**: [github.com/mnisperuza/bert-cli/issues](https://github.com/mnisperuza/bert-cli/issues)

- **Email**: mnisperuza1102@gmail.com

When reporting issues, include:
1. Bert version (`bert --ver`)
2. Your OS (Windows/Linux/macOS)
3. Error message



---
Thanks for Using Bert CLI ❤️




