Executespec/ganesh-python-v0.1.0
Ganesh Python v0.1.0 is a 12 billion parameter coding model developed by the ExecuteSpec R&D Lab, based on Gemma 4 12B IT. It is specifically focused on Python code generation, repair, and developer-assistant workflows. This model excels at tasks like code completion, debugging assistance, and refactoring, making it suitable for coding-agent and developer-tool evaluations. It offers strong performance on Python-specific benchmarks, with notable improvements over its base model on HumanEval and MBPP.
Loading preview...
Ganesh Python v0.1.0: A Specialized Python Coding Model
Ganesh Python v0.1.0 is a 12 billion parameter model developed by the ExecuteSpec R&D Lab, built upon the Gemma 4 12B IT architecture. This model is a specialized, frozen research release primarily focused on Python code generation, repair, and developer-assistant workflows. It is distributed as a standalone merged checkpoint, available in BF16 SafeTensors, GGUF Q8_0, and GGUF Q4_K_M formats.
Key Capabilities
- Python Code Generation: Generates and completes Python code.
- Code Repair & Debugging: Assists with fixing and debugging Python code.
- Refactoring & Exploration: Supports code refactoring and implementation exploration.
- Developer Tooling: Suitable for evaluating coding agents and developer tools.
Performance Highlights
On Python executable evaluations, Ganesh Python v0.1.0 showed improvements over its base model. For instance, on EvalPlus, it achieved:
- HumanEval: 96.3% (up from 95.1% base)
- HumanEval+: 93.9% (up from 92.1% base)
- MBPP: 88.4% (up from 87.8% base)
Intended Use Cases
This model is ideal for researchers and developers working with Python, particularly for tasks requiring high-fidelity code generation and assistance. It's designed for experimentation with local or hosted inference in environments like Transformers, vLLM, and Unsloth. Users should be aware that generated code requires review and testing, as it may produce incorrect or insecure outputs.