lucasjinreal / lucasjinreal/Namo-R1

Namo-V2 release

Open
#13 2 comments 8 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
256
Forks
26
PR merge metrics
No merged PRs in 30d

Description

Hello! We just finished training of Namo-500M-V2!

The v2 version boosted a lot in OCR ability, and it adopted SiglipV2 as vision encoder. Moreover, it can handle native resolution input as usual in the Namo series.

We are still conducting mDPO on the trained V2 model. We are exploring if mDPO could further enhance the tiny model's ability or not. Stay tuned for our final release.

Also, the new V2 version will have two versions, one with 578 tokens per / img, and one with 256 tokens per / img.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

The issue announces the Namo-500M-V2 model, SiglipV2 vision encoder, mDPO work, and two token configurations, but names no files, tests, or implementation entry point. It does not define a concrete change or completion condition, so further project context and maintainer direction are needed before work can start.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai, 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.