flash attn: add auto mode for llama engine (#13052)

* flash attn: add auto mode for llama engine

If the user does not specify fa in the environment, use auto-mode.

* review comments

* ensure kv cache quantized types have FA explicitly enabled

additional review comments

ml/backend.go

@@ -74,7 +74,7 @@ type BackendParams struct {
 	GPULayers GPULayersList
 
 	// FlashAttention indicates that we should use a fused flash attention kernel
-	FlashAttention bool
+	FlashAttention FlashAttentionType
 }
 
 var backends = make(map[string]func(string, BackendParams) (Backend, error))