mirrors/ollama-for-amd
1
0
Fork 0
mirror of https://github.com/likelovewant/ollama-for-amd.git synced 2025-12-21 22:33:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama: allocate grammar buffer based on schema length (#10649)

Browse Source
This commit is contained in:
frob
2025-05-10 20:57:30 +02:00
committed by GitHub
parent 69ce44b33c
commit ecf14a220f
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
llama/llama.go
View File

@@ -637,8 +637,8 @@ func SchemaToGrammar(schema []byte) []byte {
cStr := C.CString(string(schema)) cStr := C.CString(string(schema))
defer C.free(unsafe.Pointer(cStr)) defer C.free(unsafe.Pointer(cStr))
// Allocate buffer for grammar output with reasonable size // Allocate buffer for grammar based on schema length but with upper bound
const maxLen = 32768 // 32KB maxLen := min(1024*1024, len(schema)*4)
buf := make([]byte, maxLen) buf := make([]byte, maxLen)
// Call C function to convert schema to grammar // Call C function to convert schema to grammar