#gemma
6 posts tagged with #gemma
Every article below is hand-written, technically reviewed, and focused on gemma. Posts cover real-world architecture decisions, code-level implementation patterns, and trade-offs you'll only discover after shipping production systems.
AI and Machine Learning How to Evaluate EmbeddingGemma 2 for Multilingual RAG [2026]
EmbeddingGemma 2 is shiny. Shipping it into a multilingual RAG stack without regressions is not. Here’s an embeddings-only eval + rollout playbook that works.
![Fine-Tune Open-Source LLMs: LoRA, QLoRA, Gemma 4 [2026]](https://img.kunalganglani.com/images/vzekdneq/production/bbdfb150ddb4280a3411f8cbcf4d7f0f54cfdf63-1200x675.webp?auto=format&fit=max&q=75&w=600)
AI and Machine Learning Fine-Tune Open-Source LLMs: LoRA, QLoRA, Gemma 4 [2026]
A practical 2026 guide to fine-tuning open-source LLMs with LoRA and QLoRA using Unsloth + Gemma 4 — including GPU requirements, hyperparameter defaults, evaluation setup, and when to just prompt instead.
AI and Machine Learning Gemma 3 vs Llama 3 (2026): Which Open-Weight LLM Actually Wins?
Llama 3 wins for ecosystem depth, community tooling, and large-scale deployments; Gemma 3 wins for hardware efficiency, multimodal tasks, and privacy-first on-device workloads. Your hardware budget and use case should decide this — not brand loyalty.
AI and Machine Learning Run Gemma 3 Locally on Windows: The VRAM Guide Nobody Gave You [2026]
I break down exactly how much VRAM you need for every Gemma 3 model size, the Ollama setup on Windows with NVIDIA GPUs, and the performance tweaks that actually matter.
Technology Gemma 3 on Raspberry Pi 5: Benchmarked [2026]
I benchmarked every runnable Gemma variant on a Raspberry Pi 5 — Gemma 3 1B, 4B, QAT models, and Gemma 3n — with real tokens-per-second numbers, RAM usage, and a definitive size guide for Pi owners.
AI and Machine Learning Fine-Tuning Gemma 2 for Code Generation: 31 Percentage Points of Accuracy for Under $15 [2026 Guide]
I fine-tuned Google's Gemma 2 9B model for Python code generation using QLoRA on a single GPU. Accuracy jumped from 42% to 73% — here's exactly how, what it cost, and what I learned.