mirror of
https://github.com/mudler/LocalAI.git
synced 2026-11-06 12:33:16 -05:00
Ettore Di Giacinto
1834d4fd72
No release ever shipped add-ons, so there is no installed base on api 1.
The manifest must now declare api {"min": 2, "max": 2}; a manifest with
api 1, no api or a wider range is refused with a message that names the
range.
Every field that was gated on api 2 is now part of the contract, so the
Min*API constants and their gates are gone, and so are HostAPIMin and
the apiMin host field. Three older-manifest shims go with them: the
resource listing the host implied for one add-on, the legacyArgs hint
for arguments an older version of a tool took, and the tests that
checked the gates. All shipped manifests and the test fixtures declare
api 2.
Assisted-by: Claude:claude-sonnet-5-5 go test
171 lines
5.3 KiB
JSON
171 lines
5.3 KiB
JSON
{
|
|
"schemaVersion": 1,
|
|
"id": "quantization",
|
|
"version": "0.1.0",
|
|
"executable": "quantization",
|
|
"os": "linux",
|
|
"arch": "amd64",
|
|
"host": {
|
|
"min": 1,
|
|
"max": 1
|
|
},
|
|
"api": {"min": 2, "max": 2},
|
|
"protocol": {
|
|
"min": 1,
|
|
"max": 1
|
|
},
|
|
"dataFormat": {
|
|
"min": 1,
|
|
"max": 1
|
|
},
|
|
"capabilities": {
|
|
"hostOperations": {
|
|
"required": [
|
|
"quantize.get",
|
|
"quantize.import",
|
|
"quantize.list",
|
|
"quantize.start",
|
|
"quantize.stop"
|
|
]
|
|
},
|
|
"tools": [
|
|
"resource-exists",
|
|
"quantize-list",
|
|
"quantize-start",
|
|
"quantize-status",
|
|
"quantize-stop",
|
|
"quantize-import"
|
|
],
|
|
"toolDefinitions": [
|
|
{
|
|
"name": "quantize-list",
|
|
"kind": "read",
|
|
"description": "List the quantization jobs of the person you are talking to, newest first, including jobs started in the web page. A model can only control jobs it started.",
|
|
"inputSchema": {
|
|
"type": "object",
|
|
"properties": {},
|
|
"additionalProperties": false
|
|
}
|
|
},
|
|
{
|
|
"name": "quantize-start",
|
|
"kind": "create",
|
|
"description": "Start a new quantization job. In a chat it runs at once and the person can undo it; from an API key it is only a proposal that a person confirms. A model can only control jobs it started.",
|
|
"inputSchema": {
|
|
"type": "object",
|
|
"properties": {
|
|
"name": {
|
|
"type": "string",
|
|
"description": "A label for the job, 1 to 80 characters.",
|
|
"maxLength": 80
|
|
},
|
|
"model": {
|
|
"type": "string",
|
|
"description": "Model to quantize as a gallery or Hugging Face ID. A local path is refused.",
|
|
"maxLength": 256
|
|
},
|
|
"backend": {
|
|
"type": "string",
|
|
"description": "Quantization backend. The default is llama-cpp-quantization.",
|
|
"maxLength": 64
|
|
},
|
|
"quantizationType": {
|
|
"type": "string",
|
|
"description": "Quantization type, such as q4_k_m (lowercase letters, digits and underscores). The default is q4_k_m.",
|
|
"maxLength": 32
|
|
}
|
|
},
|
|
"required": [
|
|
"name",
|
|
"model"
|
|
],
|
|
"additionalProperties": false
|
|
}
|
|
},
|
|
{
|
|
"name": "quantize-status",
|
|
"kind": "read",
|
|
"description": "Show the status of a quantization job that a tool started. A model can only control jobs it started.",
|
|
"inputSchema": {
|
|
"type": "object",
|
|
"properties": {
|
|
"resourceId": {
|
|
"type": "string",
|
|
"description": "ID of the quantization job, as returned by a job start. Only a job that a tool started has one."
|
|
}
|
|
},
|
|
"required": [
|
|
"resourceId"
|
|
],
|
|
"additionalProperties": false
|
|
}
|
|
},
|
|
{
|
|
"name": "quantize-stop",
|
|
"kind": "write",
|
|
"description": "Stop a quantization job that a tool started. Stopping cannot be undone. In a chat it runs at once and the person can undo it; from an API key it is only a proposal that a person confirms. A model can only control jobs it started.",
|
|
"inputSchema": {
|
|
"type": "object",
|
|
"properties": {
|
|
"resourceId": {
|
|
"type": "string",
|
|
"description": "ID of the quantization job, as returned by a job start. Only a job that a tool started has one."
|
|
}
|
|
},
|
|
"required": [
|
|
"resourceId"
|
|
],
|
|
"additionalProperties": false
|
|
}
|
|
},
|
|
{
|
|
"name": "quantize-import",
|
|
"kind": "write",
|
|
"description": "Register the result of a quantization job that a tool started as a model. Administrators only. Importing cannot be undone. In a chat it runs at once and the person can undo it; from an API key it is only a proposal that a person confirms. A model can only control jobs it started.",
|
|
"inputSchema": {
|
|
"type": "object",
|
|
"properties": {
|
|
"resourceId": {
|
|
"type": "string",
|
|
"description": "ID of the quantization job, as returned by a job start. Only a job that a tool started has one."
|
|
},
|
|
"name": {
|
|
"type": "string",
|
|
"description": "Name of the model, 1 to 64 letters, digits, dots, dashes or underscores. Leave out to let the engine choose.",
|
|
"maxLength": 64
|
|
}
|
|
},
|
|
"required": [
|
|
"resourceId"
|
|
],
|
|
"additionalProperties": false
|
|
}
|
|
}
|
|
],
|
|
"undo": true,
|
|
"resources": {
|
|
"noun": "quantization job",
|
|
"create": "quantize-start",
|
|
"inspect": "quantize-status"
|
|
},
|
|
"ui": {
|
|
"entry": "ui/quantization.js",
|
|
"css": [
|
|
"ui/quantization.css"
|
|
],
|
|
"sdk": {
|
|
"min": 2,
|
|
"max": 2
|
|
},
|
|
"pages": [
|
|
"jobs"
|
|
],
|
|
"navigation": [],
|
|
"mainPage": "jobs",
|
|
"icon": "ui/icon.png",
|
|
"title": "Quantization",
|
|
"description": "Shrink a model to a smaller quantized file, then import it so you can chat with it."
|
|
}
|
|
}
|
|
}
|