Continuous OPD experiment handoff

Weights-only research checkpoint, not a claim of improved task correctness. See manifest.json for the exact update and SHA256.

Download opd-code.zip, extract, then read STANDALONE_HOT_RUN.md. Install pinned dependencies and obtain authorized access to the base Google models. This custom adapter is not loadable directly with AutoModel.from_pretrained. Google Gemma Terms of Use and Prohibited Use Policy apply: https://ai.google.dev/gemma/terms and https://ai.google.dev/gemma/prohibited_use_policy.

No credentials, optimizer, private prompts or run logs are included. The eight-GPU distributed OPD implementation is not yet verified.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for hydral8/diffusiongemma-continuous-opd-handoff

Base model

google/diffusiongemma-26B-A4B-it

Finetuned

(33)

this model