12 lines
438 B
Go
12 lines
438 B
Go
package llm
|
|
|
|
import (
|
|
"github.com/jmorganca/ollama/api"
|
|
)
|
|
|
|
func newDefaultExtServer(model string, adapters, projectors []string, opts api.Options) (extServer, error) {
|
|
// On windows we always load the llama.cpp libraries dynamically to avoid startup DLL dependencies
|
|
// This ensures we can update the PATH at runtime to get everything loaded
|
|
|
|
return newDynamicShimExtServer(AvailableShims["cpu"], model, adapters, projectors, opts)
|
|
}
|