File size: 7,296 Bytes
747e7be |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 |
{
"$schema": "http://json-schema.org/draft-07/schema#",
"title": "Inference Configuration",
"description": "Configuration schema for model inference entrypoint",
"type": "object",
"required": ["schema_version", "inference_type", "load_time_parameters"],
"definitions": {
"ggufFile": {
"oneOf": [
{
"type": "string",
"pattern": "^(?!/)(?:[^/\\n]+/)*[^/\\n]+\\.gguf$",
"description": "Relative path to .gguf file (no leading slash, subdirectories allowed)"
},
{
"type": "string",
"pattern": "^/(?:[^/\\n]+/)*[^/\\n]+\\.gguf$",
"description": "Absolute path to .gguf file (leading slash required)"
},
{
"type": "string",
"pattern": "^https?://[^\\s]+\\.gguf(\\?[^\\s]*)?$",
"description": "HTTPS/HTTP URL to .gguf file (query parameters allowed)"
}
]
},
"safetensorsFile": {
"oneOf": [
{
"type": "string",
"pattern": "^(?!/)(?:[^/\\n]+/)*[^/\\n]+\\.safetensors$",
"description": "Relative path to .safetensors file (no leading slash, subdirectories allowed)"
},
{
"type": "string",
"pattern": "^/(?:[^/\\n]+/)*[^/\\n]+\\.safetensors$",
"description": "Absolute path to .safetensors file (leading slash required)"
},
{
"type": "string",
"pattern": "^https?://[^\\s]+\\.safetensors(\\?[^\\s]*)?$",
"description": "HTTPS/HTTP URL to .safetensors file (query parameters allowed)"
}
]
},
"samplingParameters": {
"type": "object",
"description": "Sampling configuration for text generation",
"properties": {
"temperature": {
"type": "number",
"minimum": 0.0,
"maximum": 2.0,
"description": "Sampling temperature (0.0 = deterministic, higher = more random)"
},
"top_p": {
"type": "number",
"minimum": 0.0,
"maximum": 1.0,
"description": "Nucleus sampling probability (cumulative probability threshold)"
},
"min_p": {
"type": "number",
"minimum": 0.0,
"maximum": 1.0,
"description": "Minimum probability threshold for token consideration"
},
"repetition_penalty": {
"type": "number",
"minimum": 0.0,
"maximum": 2.0,
"description": "Penalty for token repetition (1.0 = no penalty, higher = more penalty)"
}
},
"additionalProperties": false
}
},
"properties": {
"schema_version": {
"type": "string",
"enum": ["1.0.0"],
"description": "Schema version for compatibility checking and migration"
},
"inference_type": {"type": "string", "description": "Combined inference backend and type in format: backend/type"},
"load_time_parameters": {"type": "object", "description": "Parameters required at model load time"},
"generation_time_parameters": {"type": "object", "description": "Optional parameters used during generation"}
},
"allOf": [
{
"if": {"properties": {"schema_version": {"const": "1.0.0"}}},
"then": {
"properties": {
"inference_type": {"enum": ["llama.cpp/text-to-text", "llama.cpp/image-to-text", "llama.cpp/lfm2-audio-v1"]}
},
"allOf": [
{
"if": {"properties": {"inference_type": {"const": "llama.cpp/text-to-text"}}},
"then": {
"properties": {
"load_time_parameters": {
"required": ["model"],
"properties": {
"chat_template": {"type": "string", "description": "Optional chat template override"},
"model": {"$ref": "#/definitions/ggufFile", "description": "Text model file (local path or URL)"}
},
"additionalProperties": false
},
"generation_time_parameters": {
"properties": {"sampling_parameters": {"$ref": "#/definitions/samplingParameters"}},
"additionalProperties": false
}
}
}
},
{
"if": {"properties": {"inference_type": {"const": "llama.cpp/image-to-text"}}},
"then": {
"properties": {
"load_time_parameters": {
"required": ["model", "multimodal_projector"],
"properties": {
"chat_template": {"type": "string", "description": "Optional chat template override"},
"model": {
"$ref": "#/definitions/ggufFile",
"description": "Backbone model file (local path or URL)"
},
"multimodal_projector": {
"$ref": "#/definitions/ggufFile",
"description": "Multimodal projector file (local path or URL)"
}
},
"additionalProperties": false
},
"generation_time_parameters": {
"properties": {"sampling_parameters": {"$ref": "#/definitions/samplingParameters"}},
"additionalProperties": false
}
}
}
},
{
"if": {"properties": {"inference_type": {"const": "llama.cpp/lfm2-audio-v1"}}},
"then": {
"properties": {
"load_time_parameters": {
"required": ["model", "multimodal_projector", "audio_decoder", "audio_tokenizer"],
"properties": {
"chat_template": {"type": "string", "description": "Optional chat template override"},
"model": {
"$ref": "#/definitions/ggufFile",
"description": "Backbone model file (local path or URL)"
},
"multimodal_projector": {
"$ref": "#/definitions/ggufFile",
"description": "Audio encoder file (local path or URL)"
},
"audio_decoder": {
"$ref": "#/definitions/ggufFile",
"description": "Audio decoder file (local path or URL)"
},
"audio_tokenizer": {
"$ref": "#/definitions/safetensorsFile",
"description": "Audio tokenizer file (local path or URL)"
}
},
"additionalProperties": false
},
"generation_time_parameters": {
"properties": {
"sampling_parameters": {"$ref": "#/definitions/samplingParameters"},
"number_of_decoding_threads": {
"type": "integer",
"description": "Number of threads for audio decoding",
"minimum": 1
}
},
"additionalProperties": false
}
}
}
}
]
}
}
]
}
|