File size: 2,643 Bytes
589e50e 0249933 589e50e 0249933 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 | ---
title: Forgekit
app_file: app.py
sdk: gradio
sdk_version: 5.42.0
---
# 🔥 ForgeKit
**Forge your perfect AI model — no code required.**
ForgeKit is an open-source platform that lets anyone create custom AI models by merging existing ones. No coding, no complex setup — just pick your models, configure the merge, and get a ready-to-run Colab notebook.
## ✨ Features
### ⚒️ Merge Builder
- Add models by ID and instantly check architecture compatibility
- Choose from 6 merge methods: DARE-TIES, TIES, SLERP, Linear, Task Arithmetic, Passthrough
- Adjust weights and densities with smart presets
- Auto-suggest base model and tokenizer
- Generate ready-to-run Colab notebooks with one click
### 🔍 Model Explorer
- Search HuggingFace Hub for models
- Filter by architecture type
- View detailed model specs (hidden size, layers, vocab, etc.)
### 📦 GGUF Quantizer
- Convert any HF model to GGUF format
- Multiple quantization levels (Q8_0, Q5_K_M, Q4_K_M, etc.)
- Ready-to-run Colab notebook generation
### 🚀 Deploy
- Generate deployment files for HuggingFace Spaces
- Gradio chat interface or Docker + llama.cpp options
- Auto-generated app.py and README
### 🏆 Community Leaderboard
- Browse community-created merges
- Submit your own merged models
- Discover popular merge recipes
## 🛠️ Supported Merge Methods
| Method | Models | Best For |
|--------|--------|----------|
| **DARE-TIES** | 2-10 | Combining specialists (coding + math) |
| **TIES** | 2-10 | Resolving parameter interference |
| **SLERP** | 2 | Smooth two-model interpolation |
| **Linear** | 2-10 | Simple weighted averaging |
| **Task Arithmetic** | 1-10 | Adding/removing capabilities |
| **Passthrough** | 1-10 | Layer stacking (Frankenmerge) |
## 🚀 How It Works
1. **Add Models** — Enter HuggingFace model IDs
2. **Check Compatibility** — ForgeKit verifies architectures match
3. **Configure** — Choose method, adjust weights, pick presets
4. **Generate** — Get a Colab notebook with everything pre-filled
5. **Run** — Open in Colab, click Run All, wait for your model
6. **Ship** — Auto-upload to HF Hub + optional GGUF + Space deployment
## 📋 Requirements
The generated Colab notebooks handle all dependencies. You just need:
- A Google account (for Colab)
- A HuggingFace account (for model access and upload)
- A HF token (for gated models and uploading)
## 🧑💻 Built By
**[AIencoder](https://huggingface.co/AIencoder)** — AI/ML Engineer
- [Portfolio](https://aiencoder-portfolio.static.hf.space)
- [GitHub](https://github.com/Ary5272)
## 📄 License
MIT — use it, fork it, improve it.
|