Add audio codec (untested) to huggingface runtime

[ajsalow]

All of this is untested but I think we need to add an audio codec to support sending audio bytes to hugging face ASR models.

with open("path/to/audio.wav", "rb") as f:
    audio_bytes = f.read()

audio_b64 = base64.b64encode(audio_bytes).decode()

body = {
    "inputs": [
        {
            "name": "inputs",
            "shape": [1],
            "datatype": "BYTES",
            "parameters": {
                "content_type": "audio_bytes"  # This is the key
            },
            "data": [audio_b64],  # Base64 encoded bytes
        }
    ]
}

[scarpenter44]

Tested. See notes in https://striveworks.atlassian.net/browse/FEED-1420?selectedIssue=NGC2-82

[ajsalow]

At least some of the failing tests are failing in CI because "No space left on device" changes shouldn't impact many of them, merging.