llava: build clip image from pixels (#11999)

* llava: export function `clip_build_img_from_pixels` to build image from pixels decoded by other libraries instead of stb_image.h for better performance

* Apply suggestions from code review

---------

Co-authored-by: Xuan-Son Nguyen <redacted>

diff --git a/examples/llava/clip.cpp b/examples/llava/clip.cpp
index bf70e351c9030753598dccf48f7b7f502de305e3..e2150c06d32042f926c54940df4b1e68e7205f75 100644 (file)
--- a/examples/llava/clip.cpp
+++ b/examples/llava/clip.cpp
@@ -1729,11 +1729,11 @@ void clip_image_f32_batch_free(struct clip_image_f32_batch  * batch) {
     }
 }
 
-static void build_clip_img_from_data(const stbi_uc * data, int nx, int ny, clip_image_u8 * img) {
+void clip_build_img_from_pixels(const unsigned char * rgb_pixels, int nx, int ny, clip_image_u8 * img) {
     img->nx = nx;
     img->ny = ny;
     img->buf.resize(3 * nx * ny);
-    memcpy(img->buf.data(), data, img->buf.size());
+    memcpy(img->buf.data(), rgb_pixels, img->buf.size());
 }
 
 bool clip_image_load_from_file(const char * fname, clip_image_u8 * img) {
@@ -1743,7 +1743,7 @@ bool clip_image_load_from_file(const char * fname, clip_image_u8 * img) {
         LOG_ERR("%s: failed to load image '%s'\n", __func__, fname);
         return false;
     }
-    build_clip_img_from_data(data, nx, ny, img);
+    clip_build_img_from_pixels(data, nx, ny, img);
     stbi_image_free(data);
     return true;
 }
@@ -1755,7 +1755,7 @@ bool clip_image_load_from_bytes(const unsigned char * bytes, size_t bytes_length
         LOG_ERR("%s: failed to decode image bytes\n", __func__);
         return false;
     }
-    build_clip_img_from_data(data, nx, ny, img);
+    clip_build_img_from_pixels(data, nx, ny, img);
     stbi_image_free(data);
     return true;
 }

diff --git a/examples/llava/clip.h b/examples/llava/clip.h
index 841b4f6f9098f5f9d9bdd90a01c2f012168aa8e3..f557122008ff14c22158f1f78db87612c9e73264 100644 (file)
--- a/examples/llava/clip.h
+++ b/examples/llava/clip.h
@@ -73,6 +73,9 @@ CLIP_API void clip_image_f32_free(struct clip_image_f32 * img);
 CLIP_API void clip_image_u8_batch_free (struct clip_image_u8_batch  * batch);
 CLIP_API void clip_image_f32_batch_free(struct clip_image_f32_batch * batch);
 
+/** build image from pixels decoded by other libraries instead of stb_image.h for better performance. The memory layout is RGBRGBRGB..., input buffer length must be 3*nx*ny bytes */
+CLIP_API void clip_build_img_from_pixels(const unsigned char * rgb_pixels, int nx, int ny, clip_image_u8 * img);
+
 CLIP_API bool clip_image_load_from_file(const char * fname, struct clip_image_u8 * img);
 
 /** interpret bytes as an image file with length bytes_length, and use the result to populate img */