86 lines
2.1 KiB
Markdown
86 lines
2.1 KiB
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
|
||
|
|
tags:
|
||
|
|
- ios
|
||
|
|
- swift
|
||
|
|
- swiftui
|
||
|
|
- xcode
|
||
|
|
- code-generation
|
||
|
|
- mobile-development
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
---
|
||
|
|
|
||
|
|
# StalkIQ iOS App Generator
|
||
|
|
|
||
|
|
A fine-tuned LLM that generates complete, production-ready SwiftUI iOS application codebases from natural language descriptions.
|
||
|
|
|
||
|
|
## What It Does
|
||
|
|
|
||
|
|
Give it an app idea, get back a full Xcode project with:
|
||
|
|
- `project.yml` (XcodeGen)
|
||
|
|
- SwiftUI views, view models, models
|
||
|
|
- `Info.plist`, `LaunchScreen.storyboard`, `Assets.xcassets`
|
||
|
|
- MVVM architecture, dark gradient UI theme
|
||
|
|
- iOS 16+ deployment target
|
||
|
|
|
||
|
|
## Usage
|
||
|
|
|
||
|
|
### With Ollama (Recommended)
|
||
|
|
|
||
|
|
```bash
|
||
|
|
# Download and create the model
|
||
|
|
echo 'FROM hf.co/stalkiq/stalkiq-ios-app-generator/stalkiq.gguf' > Modelfile
|
||
|
|
ollama create stalkiq -f Modelfile
|
||
|
|
|
||
|
|
# Generate an app
|
||
|
|
ollama run stalkiq "Create an iOS app called RecipeAI that identifies food from photos using AI"
|
||
|
|
```
|
||
|
|
|
||
|
|
### With Python (transformers)
|
||
|
|
|
||
|
|
```python
|
||
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||
|
|
|
||
|
|
model = AutoModelForCausalLM.from_pretrained("stalkiq/stalkiq-ios-app-generator")
|
||
|
|
tokenizer = AutoTokenizer.from_pretrained("stalkiq/stalkiq-ios-app-generator")
|
||
|
|
|
||
|
|
prompt = "### Instruction:\nCreate an iOS app called RecipeAI that identifies food from photos\n\n### Response:\n"
|
||
|
|
inputs = tokenizer(prompt, return_tensors="pt")
|
||
|
|
outputs = model.generate(**inputs, max_new_tokens=2048, temperature=0.7, do_sample=True)
|
||
|
|
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
|
||
|
|
```
|
||
|
|
|
||
|
|
### API
|
||
|
|
|
||
|
|
```bash
|
||
|
|
# Via Ollama API (after creating the model)
|
||
|
|
curl http://localhost:11434/api/generate -d '{
|
||
|
|
"model": "stalkiq",
|
||
|
|
"prompt": "Create an iOS app called WeatherAI that shows forecasts from sky photos"
|
||
|
|
}'
|
||
|
|
```
|
||
|
|
|
||
|
|
## Prompt Format
|
||
|
|
|
||
|
|
```
|
||
|
|
### Instruction:
|
||
|
|
{your app description here}
|
||
|
|
|
||
|
|
### Response:
|
||
|
|
```
|
||
|
|
|
||
|
|
## Model Details
|
||
|
|
|
||
|
|
- **Base Model**: TinyLlama-1.1B-Chat-v1.0
|
||
|
|
- **Fine-tuning**: QLoRA on StalkIQ iOS app training data
|
||
|
|
- **Architecture**: LlamaForCausalLM (1.1B parameters)
|
||
|
|
- **Tensor Type**: FP16
|
||
|
|
- **Files**: safetensors (for transformers) + GGUF (for Ollama/llama.cpp)
|
||
|
|
|
||
|
|
## Built By
|
||
|
|
|
||
|
|
[StalkIQ](https://stalkiq.dev) — AI-powered web and mobile applications.
|