karpathy / karpathy/llm.c

image-gpt

Open
#697 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Cuda
Stars
31k
Forks
3.8k
PR merge metrics
No merged PRs in 30d

Description

@karpathy Please add code for training of 125M [image-gpt](https://github.com/openai/image-gpt) to this repository. May be, also extend the context length to 2k so that we can get 45x45 pixel image instead of 32x32 pixel and also make the color palette R4G4B4 instead of R3G3B3 for crispier image. Also, I guess a recent dataset like LAION could be used. Then we will a get a small text-to image generator which although much underpowered, will be useful in many cases.

Contributor guide

No contributing guide indexed for this repository

Research direction

No files, tests, or entry points are named. Start by locating the repository's existing training entry points and comparing them with the requested 125M image-GPT training, 2k context, R4G4B4 palette, and LAION data. Done would require an agreed scope plus working training code and validation for the requested image-generation changes.

Written by the indexing model from the issue text.

Assessment

Domain
ai, computer-vision, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
15/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.