-
Couldn't load subscription status.
- Fork 13.4k
Open
Labels
Description
Name and Version
$ llama-cli --version
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 3090, compute capability 8.6, VMM: yes
version: 6841 (54c618222)
built with MSVC 19.43.34810.0 for x64
Original issue by @serjrd #16714 (comment)
Operating systems
Windows
Which llama.cpp modules do you know to be affected?
llama-server
Command line
Problem description & steps to reproduce
The JSON schema to grammar does not support references to items in an array.
curl http://localhost:8080/v1/chat/completions -d '{
"model": "gpt-oss-20b",
"messages": [
{
"role": "user",
"content": "Build a binary tree that matches (A (B C D) E)"
}
],
"tools": [
{
"type": "function",
"function": {
"name": "build_binary_tree",
"description": "build a binary tree. The left/right fields may be either a string value or another tree.",
"parameters": {
"type": "object",
"properties": {
"tree": {
"type": "object",
"properties": {
"left": {
"anyOf": [
{
"type": "string"
},
{
"$ref": "#/properties/tree"
}
]
},
"right": {
"anyOf": [
{
"$ref": "#/properties/tree/properties/left/anyOf/0"
},
{
"$ref": "#/properties/tree"
}
]
}
},
"additionalProperties": false
}
},
"required": [
"tree"
],
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}
}
}
],
"tool_choice": "auto"
}'
First Bad Commit
No response
Relevant log output
{
"error": {
"code": 500,
"message": "JSON schema conversion failed:\nError resolving ref #/properties/tree/properties/left/anyOf/0: 0 not in [{\"type\":\"string\"},{\"$ref\":\"#/properties/tree\"}]",
"type": "server_error"
}
}