server: (router) expose child model info from router's /v1/models

[ngxson]

Overview

Allow server router instance to reflect the child instance's /v1/models info in the list of all models. This is possible for loaded models only.

For example:

{
    "data": [
        {
            "id": "ggml-org/GLM-OCR-GGUF:Q8_0",
            "aliases": [],
            "tags": [],
            "object": "model",
            "owned_by": "llamacpp",
            "created": 1777909820,
            "status": {
                "value": "loaded",
                "args": [
                    ....
                ],
                "preset": "[ggml-org/GLM-OCR-GGUF:Q8_0]\nhf-repo = ggml-org/GLM-OCR-GGUF:Q8_0\n\n"
            },
            "meta": {
                "vocab_type": 2,
                "n_vocab": 59392,
                "n_ctx": 131072,
                "n_ctx_train": 131072,
                "n_embd": 1536,
                "n_params": 891138048,
                "size": 947159040
            }
        }
    }
}

Requirements

• I have read and agree with the contributing guidelines
• AI usage disclosure: no

[ggerganov]

Better allocate this on the heap