CodeBooga-34B-v0.1-GGUF
| Property | Value |
|---|---|
| Parameter Count | 33.7B |
| Model Type | LLaMA Architecture |
| License | LLaMA 2 |
| Base Models | Phind-CodeLlama-34B-v2 + WizardCoder-Python-34B-V1.0 |
What is CodeBooga-34B-v0.1-GGUF?
CodeBooga-34B is an advanced code generation model created through a sophisticated merge of Phind-CodeLlama-34B-v2 and WizardCoder-Python-34B-V1.0. This GGUF version offers various quantization options for efficient deployment, ranging from 2-bit to 8-bit precision, making it adaptable to different hardware configurations and performance requirements.
Implementation Details
The model utilizes the BlockMerge Gradient script for combining the base models, with carefully calibrated gradient values for different components. It implements the Alpaca prompt format and offers multiple quantization options through GGUF, including Q2_K through Q8_0 variants, each optimized for different use-cases and hardware constraints.
- Multiple quantization options (2-bit to 8-bit)
- Optimized for both Python and JavaScript code generation
- Implements Alpaca prompt format for consistent interaction
- Supports GPU acceleration with layer offloading
Core Capabilities
- Advanced code generation and completion
- Multi-language programming support
- Flexible deployment options with various quantization levels
- Extended context handling up to 4096 tokens
Frequently Asked Questions
Q: What makes this model unique?
The model combines the strengths of both Phind-CodeLlama and WizardCoder, offering superior performance in code generation tasks. In comparative evaluations, it achieved significantly higher scores (22 points) compared to its base models (12 and 7 points respectively).
Q: What are the recommended use cases?
CodeBooga-34B is ideal for code generation, completion, and programming assistance tasks, particularly in Python and JavaScript. It's recommended to use with the Divine Intellect preset settings for optimal performance in instruction-following scenarios.
