ollamarunner: Pass runner performance parameters to backends

Currently the following parameters are in the runner but not used:
 - numGPULayers
 - mainGPU
 - threads
 - tensorSplit

This passes them through to the backend, which is where they would
actually get used. However, the GGML backend does not yet do anything
with them.

ml/backend/ggml/ggml.go

@@ -84,7 +84,7 @@ type Backend struct {
 	tensors    map[string]*Context
 }
 
-func New(r *os.File) (ml.Backend, error) {
+func New(r *os.File, params ml.BackendParams) (ml.Backend, error) {
 	meta, n, err := fs.Decode(r, -1)
 	if err != nil {
 		return nil, err