The Kaitchup – AI on a Budget

The Kaitchup – AI on a Budget

Home
Notes
Start Here
AI Notebooks
The Kaitchup's Book
Weekly Kaitchup
Tutorials
The Kaitchup Index
Archive
About

Tutorials

Why Increasing Batch Size Doesn’t Always Speed Up Training
5 most common issues that decreases the batch training efficiency
 • 
Benjamin Marie
Serve Multiple LoRA Adapters with vLLM and Custom Chat Templates
Swap adapters per request, reuse your chat template, and run offline or via an OpenAI-compatible server.
 • 
Benjamin Marie
DenseMixer: Smarter MoE Routing That Doesn’t Break LoRA and QLoRA
Better MoE training for a slightly higher cost
 • 
Benjamin Marie
2
Gemma 3 270M: Can Tiny Models Learn New Tasks?
A case study with machine translation
 • 
Benjamin Marie
10
NVFP4: Same Accuracy with 2.3x Higher Throughput for 4-Bit LLMs
How to quantize LLMs with NVFP4
 • 
Benjamin Marie
6
How to Run Unsloth on Multi-GPU Setups: Data-Parallel or Model-Parallel
Step-by-step fixes for running Unsloth across GPUs
 • 
Benjamin Marie
3
Gemma 3n: Fine-Tuning, Inference, and Submodel Extraction
Running Gemma 3n with vLLM and fine-tuning with TRL
 • 
Benjamin Marie
RAG with Qwen3 Embedding and Qwen3 Reranker
How to use embedding and reranker models to efficiently retrieve only the most relevant chunks or documents given a user query
 • 
Benjamin Marie
4
RTX 6000 Pro vs H100 & A100: Best Single-GPU Choice for Fast, Low-Cost LLM Fine-Tuning
Faster, cheaper single-GPU training
 • 
Benjamin Marie
1
Fine-Tuning 2-Bit Qwen3 Models on Your Computer
Code and best practices
 • 
Benjamin Marie
Qwulu 3: Fine-Tuning Qwen3 Base with LoRA and TULU 3's Supervised Fine-Tuning Recipe
Can a supervised fine-tuning recipe that works effectively on Llama 3.1 be applied directly to Qwen3?
 • 
Benjamin Marie
Boost 2-Bit LLM Accuracy with EoRA
A training-free solution for extreme LLM compression
 • 
Benjamin Marie
© 2025 The Kaitchup
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture