Replace Cached with Yolox-graphic-elements and drop Deplot

[edknv]

Description

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[randerzander]

Let's also go ahead and make the switch to the llama embedder. This is what I was testing on the single GPU deployment.

EMBEDDING_IMAGE=nvcr.io/nim/nvidia/llama-3.2-nv-embedqa-1b-v2
EMBEDDING_TAG=latest
EMBEDDING_NIM_MODEL_NAME=nvidia/llama-3.2-nv-embedqa-1b-v2

[edknv]

Switched to llama in 41f9e7b.

[randerzander]

Please also add to top of README:

**NOTICE**:
Cached and Deplot are deprecated, docker-compose now points to a beta version of the yolox-graphic-elements container instead. That model and container is slated for full release in March.

With this change, you should now be able to run on a single 80GB A100 or H100 GPU.

and pre-requisites, move number of A100s and H100s to 1x

[edknv]

Noitce added at the top of REAMDE and minimum requirements adjusted to 1 GPU.
fa6e750
fd65f2d
20c290f