AIQB
TutorialsOrdinary

My first fine-tuning patch made my model worse. The second one, with fewer examples and half the learning rate, fixed it

Source: DEV Community·

Summary

Notes from shipping v5.1 of a 7B model that designs pixel art games: what went wrong with the first patch, what worked, the numbers, and why the project is now pixel art only.
CategoryAI Tutorials & Practice
TierOrdinary
Published
Indexed by AIQB
SourceDEV Community
AIQB record IDintel-792b7de3d50852fdce90c856
Read original source ↗More Tutorials

Related stories

Matched by category, entity tags and source

TutorialsOrdinary
LazyReader - a TL;DR machine that runs entirely in your browser (no servers, no keys, no excuses)
Disclosure: this project and post were created with AI assistance - the code was written by an AI...
Source: DEV Community
TutorialsOrdinary
DuckDB 2.0's agent mode is a footer, and -json throws it away
The first time I ran DuckDB 2.0 today, it printed one line before my query: duckdb agent mode on:...
Source: DEV Community
TutorialsOrdinary
SkillSentry — The Laziest Way to Keep Your AI Agent's Skills Honest
This is a submission for the Hacktoberfest Week 3 DEV Challenge Prize Categories: Best Use of...
Source: DEV Community
TutorialsOrdinary
I Fixed WebGPU on Linux. Then Forgot How I Did It.
I spent a ridiculous amount of time trying to get WebGPU working properly on my Linux machine. I...
Source: DEV Community
TutorialsOrdinary
GAI: Let AI Generate Your Git Commit Messages Locally
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady. ...
Source: DEV Community
TutorialsOrdinary
Why We Built a Self-Improving AI Coding Runtime Using Only Python's Standard Library (Zero Dependencies)
Runesmith is an open-source runtime that wraps whatever AI model you have (a free Gemini or Groq key,...
Source: DEV Community · Gemini
TutorialsOrdinary
Google Clarifies Search and AI Visibility Controls for Website Owners
Google has updated its guidance for website owners on controlling how content appears in Google...
Source: DEV Community
TutorialsOrdinary
How to Get Historical Stock Prices via Search API
Day 21 of 30 Days of Search. TL;DR To get historical stock prices with Valyu Search API,...
Source: DEV Community
TutorialsOrdinary
39 screenshots and not one useful name, so a local Gemma names them for me
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady ...
Source: DEV Community
TutorialsOrdinary
Who Watches the Watchers? I Built a Tamper-Evident Seal for Agent Guardrails
Today OpenAI disclosed something that should make every agent builder uncomfortable: its own AI...
Source: DEV Community · OpenAI