feat: implement ESRGAN upscaler + Metal Backend (#104)

* add esrgan upscaler

* add sd_tiling

* support metal backend

* add clip_skip

---------

Co-authored-by: leejet <leejet714@gmail.com>

CMakeLists.txt

@@ -25,7 +25,9 @@ endif()
 #option(SD_BUILD_TESTS                "sd: build tests"    ${SD_STANDALONE})
 option(SD_BUILD_EXAMPLES             "sd: build examples" ${SD_STANDALONE})
 option(SD_CUBLAS                     "sd: cuda backend" OFF)
+option(SD_METAL                      "sd: metal backend" OFF)
 option(SD_FLASH_ATTN                 "sd: use flash attention for x4 less memory usage" OFF)
+option(SD_FAST_SOFTMAX               "sd: x1.5 faster softmax, indeterministic (sometimes, same seed don't generate same image), cuda only" OFF)
 option(BUILD_SHARED_LIBS             "sd: build shared libs" OFF)
 #option(SD_BUILD_SERVER               "sd: build server example"                           ON)
 
@@ -33,6 +35,15 @@ if(SD_CUBLAS)
 	message("Use CUBLAS as backend stable-diffusion")
     set(GGML_CUBLAS ON)
     add_definitions(-DSD_USE_CUBLAS)
+    if(SD_FAST_SOFTMAX)
+        set(GGML_CUDA_FAST_SOFTMAX ON)
+    endif()
+endif()
+
+if(SD_METAL)
+	message("Use Metal as backend stable-diffusion")
+    set(GGML_METAL ON)
+    add_definitions(-DSD_USE_METAL)
 endif()
 
 if(SD_FLASH_ATTN)