security-coding-gemma4-e2b — Gemma 4 Security Coding Model

What Is This?

Uploaded to Hugging Face by k1ngtai. The security-coding-gemma4-e2b family is a Gemma 4 fine-tune positioned for secure coding assistance, with three related artifacts:

• security-coding-gemma4-e2b-full — full safetensors weights
• security-coding-gemma4-e2b-lora — LoRA adapter
• security-coding-gemma4-e2b-gguf — quantized inference build

Metadata

Field	Value
Author	k1ngtai
Architecture	Gemma 4 (image-text-to-text pipeline tag)
Training stack	Unsloth, TRL
License	Apache 2.0
Last modified	2026-06-15

Why It Might Matter

Code-security fine-tunes on new architecture generations (Gemma 4) are early indicators of where the community invests training compute:

• Secure coding vs. vulnerability finding — unclear from tags alone; warrants benchmark on CVEfixes-style datasets
• E2B naming — may reference sandboxed code execution environments; check model card for training methodology
• Multimodal pipeline tag — if vision inputs were used, attack surface for prompt injection via screenshots/docs expands

Comparison Targets

Benchmark against:

• Younis2003/CodeLlama_for_code_security (CodeLlama 13B LoRA)
• Qwen2.5-Coder family (general code, not security-specific)
• Microsoft/CodeBERT security classifiers

Attribution

Open-source model indexed for security research inquiry. Uploaded to Hugging Face by k1ngtai.