cleanup: do not force VAE type to f32 on SDXL

This seems to be a leftover from the initial SDXL support: it's
not enough to avoid NaN issues, and it's not not needed for the
fixed sdxl-vae-fp16-fix .

Author: wbruna

stable-diffusion.cpp

@@ -269,11 +269,6 @@ class StableDiffusionGGML {
             model_loader.set_wtype_override(wtype);
         }
 
-        if (sd_version_is_sdxl(version)) {
-            vae_wtype = GGML_TYPE_F32;
-            model_loader.set_wtype_override(GGML_TYPE_F32, "vae.");
-        }
-
         LOG_INFO("Weight type:                 %s", model_wtype != GGML_TYPE_COUNT ? ggml_type_name(model_wtype) : "??");
         LOG_INFO("Conditioner weight type:     %s", conditioner_wtype != GGML_TYPE_COUNT ? ggml_type_name(conditioner_wtype) : "??");
         LOG_INFO("Diffusion model weight type: %s", diffusion_model_wtype != GGML_TYPE_COUNT ? ggml_type_name(diffusion_model_wtype) : "??");