Confucius4-R2T2 AI Pod runtimes

ARM64 Docker load archives for AGX Thor and AGX Orin. These archives contain the inference runtime and dependencies. Confucius4-R2T2 model weights are downloaded separately from netease-youdao/Confucius4-R2T2 and mounted read-only.

Device	Archive	Bytes	SHA256
thor	`r2t2-thor-0.1.8.tar.gz`	682650204	`d6bcbc7dfa660515e3619193f99e8149633ce094e3a4c3cc743b4fb3b0fd1477`
orin	`r2t2-orin-0.1.8.tar.gz`	634926285	`d8eced6db9fe5bc0ea407248aead9fdd47378f1dbec24c7e12fa775181d13ff8`

Load with docker load -i <archive>. Verify size, SHA256 and image identity using checksums.json. Thor uses one worker and Orin uses four workers. AI Pod source is Apache-2.0; third-party dependencies remain subject to their own licenses and notices in the images. MODEL_LICENSE records the separately downloaded model's terms.

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