cleanup and refactor *again*

examples/llava/llava-cli.cpp

@@ -20,46 +20,47 @@ static void show_additional_info(int /*argc*/, char ** argv) {
     printf("  note: a lower temperature value like 0.1 is recommended for better quality.\n");
 }
 
-static bool load_image(llava_context * ctx_llava, gpt_params * params, float **image_embd, int * n_img_pos) {
+static struct llava_image_embed * load_image(llava_context * ctx_llava, gpt_params * params) {
+
     // load and preprocess the image
-    clip_image_u8 * img = make_clip_image_u8();
+    llava_image_embed * embed = NULL;
     auto prompt = params->prompt;
     if (prompt_contains_image(prompt)) {
         if (!params->image.empty()) {
             printf("using base64 encoded image instead of command line image path\n");
         }
-        if (!clip_image_load_from_prompt(prompt, img)) {
+        embed = llava_image_embed_make_with_prompt_base64(ctx_llava->ctx_clip, params->n_threads, prompt);
+        if (!embed) {
             fprintf(stderr, "%s: can't load image from prompt\n", __func__);
-            return false;
+            return NULL;
         }
         params->prompt = remove_image_from_prompt(prompt);
     } else {
-        if (!clip_image_load_from_file(params->image.c_str(), img)) {
+        embed = llava_image_embed_make_with_filename(ctx_llava->ctx_clip, params->n_threads, params->image.c_str());
+        if (!embed) {
             fprintf(stderr, "%s: is %s really an image file?\n", __func__, params->image.c_str());
-            return false;
+            return NULL;
         }
     }
-    bool image_embed_result = llava_build_img_embed(ctx_llava->ctx_llama, ctx_llava->ctx_clip, params->n_threads, img, image_embd, n_img_pos);
-    if (!image_embed_result) {
-        clip_image_u8_free(img);
-        fprintf(stderr, "%s: coulnd't embed the image\n", __func__);
-        return false;
-    }
 
-    return true;
+    return embed;
 }
 
-static void process_prompt(struct llava_context * ctx_llava, float * image_embd, int n_img_pos, gpt_params * params, const char * prompt) {
+static void process_prompt(struct llava_context * ctx_llava, struct llava_image_embed * image_embed, gpt_params * params, const char * prompt) {
     int n_past = 0;
 
     const int max_tgt_len = params->n_predict < 0 ? 256 : params->n_predict;
 
     // llava chat format is "<system_prompt>USER: <image_embeddings>\n<textual_prompt>\nASSISTANT:"
     // GG: are we sure that the should be a trailing whitespace at the end of this string?
+    printf("evaluating system prompt\n");
     eval_string(ctx_llava->ctx_llama, "A chat between a curious human and an artificial intelligence assistant.  The assistant gives helpful, detailed, and polite answers to the human's questions.\nUSER: ", params->n_batch, &n_past);
-    llava_eval_image_embd(ctx_llava->ctx_llama, image_embd, n_img_pos, params->n_batch, &n_past);
+    printf("evaluating image embed\n");
+    llava_eval_image_embed(ctx_llava->ctx_llama, image_embed, params->n_batch, &n_past);
+    printf("evaluating prompt\n");
     eval_string(ctx_llava->ctx_llama, prompt, params->n_batch, &n_past);
     eval_string(ctx_llava->ctx_llama, "\nASSISTANT:",        params->n_batch, &n_past);
+    printf("awaiting response\n");
 
     // generate the response
 
@@ -153,16 +154,14 @@ int main(int argc, char ** argv) {
         return 1;
     }
 
-    float * image_embd;
-    int n_image_pos;
-    load_image(ctx_llava, &params, &image_embd, &n_image_pos);
+    auto image_embed = load_image(ctx_llava, &params);
 
     // process the prompt
-    process_prompt(ctx_llava, image_embd, n_image_pos, &params, params.prompt.c_str());
+    process_prompt(ctx_llava, image_embed, &params, params.prompt.c_str());
 
     llama_print_timings(ctx_llava->ctx_llama);
 
-    free(image_embd);
+    llava_image_embed_free(image_embed);
     llava_free(ctx_llava);
     return 0;
 }

ggml-metal.m

@@ -189,6 +189,7 @@ static void ggml_metal_log(enum ggml_log_level level, const char* format, ...){
     {
         NSBundle * bundle = nil;
 #ifdef SWIFT_PACKAGE
+        print("would use SWIFTPM_MODULE_BUNDLE");
         bundle = SWIFTPM_MODULE_BUNDLE;
 #else
         bundle = [NSBundle bundleForClass:[GGMLMetalClass class]];

llava/clip.cpp

@@ -678,7 +678,10 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) {
     return new_clip;
 }
 
-clip_image_u8 * make_clip_image_u8() { return new clip_image_u8(); }
+clip_image_u8 * make_clip_image_u8() { 
+    auto img = new clip_image_u8();
+    return img;
+}
 clip_image_f32 * make_clip_image_f32() { return new clip_image_f32(); }
 
 void clip_image_u8_free(clip_image_u8 * img) { if (img->data) { delete[] img->data; } delete img; }
@@ -692,31 +695,30 @@ static void build_clip_img_from_data(const stbi_uc * data, int nx, int ny, clip_
     memcpy(img->data, data, img->size);
 }
 
-bool clip_image_load_from_bytes(const unsigned char * bytes, size_t bytes_length, clip_image_u8 * img) {
+bool clip_image_load_from_file(const char * fname, clip_image_u8 * img) {
     int nx, ny, nc;
-    auto data = stbi_load_from_memory(bytes, bytes_length, &nx, &ny, &nc, 3);
+    auto data = stbi_load(fname, &nx, &ny, &nc, 3);
     if (!data) {
-        fprintf(stderr, "%s: failed to decode image bytes\n", __func__);
+        fprintf(stderr, "%s: failed to load image '%s'\n", __func__, fname);
         return false;
     }
     build_clip_img_from_data(data, nx, ny, img);
     stbi_image_free(data);
     return true;
 }
 
-bool clip_image_load_from_file(const char * fname, clip_image_u8 * img) {
+bool clip_image_load_from_bytes(const unsigned char * bytes, size_t bytes_length, struct clip_image_u8 * img) {
     int nx, ny, nc;
-    auto data = stbi_load(fname, &nx, &ny, &nc, 3);
+    auto data = stbi_load_from_memory(bytes, bytes_length, &nx, &ny, &nc, 3);
     if (!data) {
-        fprintf(stderr, "%s: failed to load image '%s'\n", __func__, fname);
+        fprintf(stderr, "%s: failed to decode image bytes\n", __func__);
         return false;
     }
     build_clip_img_from_data(data, nx, ny, img);
     stbi_image_free(data);
     return true;
 }
 
-
 // normalize: x = (x - mean) / std
 // TODO: implement bicubic interpolation instead of linear.
 bool clip_image_preprocess(const clip_ctx * ctx, const clip_image_u8 * img, clip_image_f32 * res, const bool pad2square) {
@@ -1065,16 +1067,16 @@ bool clip_model_quantize(const char * fname_inp, const char * fname_out, const i
     return true;
 }
 
-int clip_n_mmproj_embd(struct clip_ctx * ctx) {
+int clip_n_mmproj_embd(const struct clip_ctx * ctx) {
     return ctx->vision_model.mm_2_b->ne[0];
 }
 
-int clip_n_patches(struct clip_ctx * ctx) {
+int clip_n_patches(const struct clip_ctx * ctx) {
     auto & params = ctx->vision_model.hparams;
 
     return (params.image_size / params.patch_size) * (params.image_size / params.patch_size);
 }
 
-size_t clip_embd_nbytes(struct clip_ctx * ctx) {
+size_t clip_embd_nbytes(const struct clip_ctx * ctx) {
     return clip_n_patches(ctx) * clip_n_mmproj_embd(ctx) * sizeof(float);
 }

llava/clip.h

@@ -25,9 +25,9 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity);
 
 void clip_free(struct clip_ctx * ctx);
 
-size_t clip_embd_nbytes(struct clip_ctx * ctx);
-int clip_n_patches(struct clip_ctx * ctx);
-int clip_n_mmproj_embd(struct clip_ctx * ctx);
+size_t clip_embd_nbytes(const struct clip_ctx * ctx);
+int clip_n_patches(const struct clip_ctx * ctx);
+int clip_n_mmproj_embd(const struct clip_ctx * ctx);
 
 // RGB uint8 image
 struct clip_image_u8 {
@@ -62,7 +62,7 @@ LLAMA_API void clip_image_u8_free(clip_image_u8 * img);
 LLAMA_API void clip_image_f32_free(clip_image_f32 * img);
 LLAMA_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 */
-LLAMA_API bool clip_image_load_from_bytes(const unsigned char * bytes, size_t bytes_length, clip_image_u8 * img);
+LLAMA_API bool clip_image_load_from_bytes(const unsigned char * bytes, size_t bytes_length, struct clip_image_u8 * img);
 
 bool clip_image_preprocess(const struct clip_ctx * ctx, const struct clip_image_u8 * img, struct clip_image_f32 * res, const bool pad2square);
 bool clip_image_encode(const struct clip_ctx * ctx, const int n_threads, struct clip_image_f32 * img, float * vec);

llava/llava-utils.h

@@ -4,6 +4,7 @@
 
 #include "common.h"
 #include "llama.h"
+#include "llava.h"
 
 #include "base64.hpp"
 
@@ -143,12 +144,12 @@ inline bool prompt_contains_image(const std::string& prompt) {
 }
 
 // replaces the base64 image tag in the prompt with `replacement`
-inline bool clip_image_load_from_prompt(const std::string& prompt, clip_image_u8 * img) {    
+inline llava_image_embed * llava_image_embed_make_with_prompt_base64(struct clip_ctx * ctx_clip, int n_threads, const std::string& prompt) {    
     size_t img_base64_str_start, img_base64_str_end;
     find_image_tag_in_prompt(prompt, img_base64_str_start, img_base64_str_end);
     if (img_base64_str_start == std::string::npos || img_base64_str_end == std::string::npos) {
         fprintf(stderr, "%s: invalid base64 image tag. must be %s<base64 byte string>%s\n", __func__, IMG_BASE64_TAG_BEGIN, IMG_BASE64_TAG_END);
-        return false;
+        return NULL;
     }
 
     auto base64_bytes_start = img_base64_str_start + strlen(IMG_BASE64_TAG_BEGIN);
@@ -157,16 +158,15 @@ inline bool clip_image_load_from_prompt(const std::string& prompt, clip_image_u8
 
     auto required_bytes = base64::required_encode_size(base64_str.size());
     auto img_bytes = std::vector<unsigned char>(required_bytes);
-    auto img_bytes_end = base64::decode(base64_str.begin(), base64_str.end(), img_bytes.begin());
-    size_t img_bytes_len = img_bytes_end - img_bytes.begin();
+    base64::decode(base64_str.begin(), base64_str.end(), img_bytes.begin());
 
-    auto img_loaded_ok = clip_image_load_from_bytes(img_bytes.data(), img_bytes_len, img);
-    if (!img_loaded_ok) {
+    auto embed = llava_image_embed_make_with_bytes(ctx_clip, n_threads, img_bytes.data(), img_bytes.size());
+    if (!embed) {
         fprintf(stderr, "%s: could not load image from base64 string.\n", __func__);
-        return false;
+        return NULL;
     }
 
-    return true;
+    return embed;
 }
 
 inline std::string remove_image_from_prompt(const std::string& prompt, const char * replacement = "") {

llava/llava.cpp

@@ -10,7 +10,7 @@
 
 #include "base64.hpp"
 
-static bool encode_image_with_clip(clip_ctx * ctx_clip, int n_threads, const clip_image_u8 * img, float * image_embd, int * n_image_embd, int * n_img_pos) {
+static bool encode_image_with_clip(clip_ctx * ctx_clip, int n_threads, const clip_image_u8 * img, float * image_embd, int * n_img_pos) {
     clip_image_f32 * img_res = make_clip_image_f32();
     if (!clip_image_preprocess(ctx_clip, img, img_res, /*pad2square =*/ true)) {
         fprintf(stderr, "%s: unable to preprocess image\n", __func__);
@@ -19,7 +19,6 @@ static bool encode_image_with_clip(clip_ctx * ctx_clip, int n_threads, const cli
     }
 
     *n_img_pos = clip_n_patches(ctx_clip);
-    *n_image_embd = clip_n_mmproj_embd(ctx_clip);
 
     const int64_t t_img_enc_start_us = ggml_time_us();
     bool encoded = clip_image_encode(ctx_clip, n_threads, img_res, image_embd);
@@ -39,7 +38,18 @@ static bool encode_image_with_clip(clip_ctx * ctx_clip, int n_threads, const cli
     return true;
 }
 
-bool llava_build_img_embed(const llama_context * ctx_llama, clip_ctx * ctx_clip, int n_threads, const clip_image_u8 * img, float ** image_embd_out, int * n_img_pos_out) {
+bool llava_validate_embed_size(const llama_context * ctx_llama, const clip_ctx * ctx_clip) {
+        // make sure that the correct mmproj was used, i.e., compare apples to apples
+    int n_llama_embd = llama_n_embd(llama_get_model(ctx_llama));
+    auto n_image_embd = clip_n_mmproj_embd(ctx_clip);
+    if (n_image_embd != n_llama_embd) {
+        printf("%s: embedding dim of the multimodal projector (%d) is not equal to that of LLaMA (%d). Make sure that you use the correct mmproj file.\n", __func__, n_image_embd, n_llama_embd);
+        return false;
+    }
+    return true;
+}
+
+static bool llava_image_embed_make_with_clip_img(clip_ctx * ctx_clip, int n_threads, const clip_image_u8 * img, float ** image_embd_out, int * n_img_pos_out) {
 
     float * image_embd = (float *)malloc(clip_embd_nbytes(ctx_clip));
     if (!image_embd) {
@@ -49,20 +59,11 @@ bool llava_build_img_embed(const llama_context * ctx_llama, clip_ctx * ctx_clip,
     }
 
     int n_img_pos;
-    int n_image_embd;
-    if (!encode_image_with_clip(ctx_clip, n_threads, img, image_embd, &n_image_embd, &n_img_pos)) {
+    if (!encode_image_with_clip(ctx_clip, n_threads, img, image_embd, &n_img_pos)) {
         fprintf(stderr, "%s: cannot encode image, aborting\n", __func__);
         free(image_embd);
         return false;
     }
-    // make sure that the correct mmproj was used, i.e., compare apples to apples
-    int n_llama_embd = llama_n_embd(llama_get_model(ctx_llama));
-    if (n_image_embd != n_llama_embd) {
-        printf("%s: embedding dim of the multimodal projector (%d) is not equal to that of LLaMA (%d). Make sure that you use the correct mmproj file.\n", __func__, n_image_embd, n_llama_embd);
-        free(image_embd);
-        return false;
-    }
-
     *image_embd_out = image_embd;
     *n_img_pos_out = n_img_pos;
 
@@ -71,15 +72,15 @@ bool llava_build_img_embed(const llama_context * ctx_llama, clip_ctx * ctx_clip,
 
 
 
-bool llava_eval_image_embd(llama_context * ctx_llama, float * image_embd, int n_image_pos, int n_batch, int * n_past) {
+bool llava_eval_image_embed(llama_context * ctx_llama, const struct llava_image_embed * image_embed, int n_batch, int * n_past) {
     int n_embd  = llama_n_embd(llama_get_model(ctx_llama));
 
-    for (int i = 0; i < n_image_pos; i += n_batch) {
-        int n_eval = n_image_pos - i;
+    for (int i = 0; i < image_embed->n_image_pos; i += n_batch) {
+        int n_eval = image_embed->n_image_pos - i;
         if (n_eval > n_batch) {
             n_eval = n_batch;
         }
-        llama_batch batch = {int32_t(n_eval), nullptr, (image_embd+i*n_embd), nullptr, nullptr, nullptr, *n_past, 1, 0, };
+        llama_batch batch = {int32_t(n_eval), nullptr, (image_embed->embed+i*n_embd), nullptr, nullptr, nullptr, *n_past, 1, 0, };
         if (llama_decode(ctx_llama, batch)) {
             fprintf(stderr, "%s : failed to eval\n", __func__);
             return false;
@@ -88,3 +89,79 @@ bool llava_eval_image_embd(llama_context * ctx_llama, float * image_embd, int n_
     }
     return true;
 }
+
+
+LLAMA_API struct llava_image_embed * llava_image_embed_make_with_bytes(struct clip_ctx * ctx_clip, int n_threads, const unsigned char * image_bytes, int image_bytes_length)
+{
+    clip_image_u8 * img = make_clip_image_u8();
+    if (!clip_image_load_from_bytes(image_bytes, image_bytes_length, img)) {
+        clip_image_u8_free(img);
+        fprintf(stderr, "%s: can't load image from bytes, is it a valid image?", __func__);
+        return NULL;
+    }
+
+    float* image_embed = NULL;
+    int n_image_pos = 0;
+    bool image_embed_result = llava_image_embed_make_with_clip_img(ctx_clip, n_threads, img, &image_embed, &n_image_pos);
+    if (!image_embed_result) {
+        clip_image_u8_free(img);
+        fprintf(stderr, "%s: coulnd't embed the image\n", __func__);
+        return NULL;
+    }
+
+    clip_image_u8_free(img);
+    auto result = (llava_image_embed*)malloc(sizeof(llava_image_embed));
+    result->embed = image_embed;
+    result->n_image_pos = n_image_pos;
+    return result;
+}
+
+static bool load_file_to_bytes(const char* path, unsigned char** bytesOut, long *sizeOut) 
+{
+    auto file = fopen(path, "rb"); 
+    if (file == NULL) {
+        fprintf(stderr, "%s: can't read file %s\n", __func__, path);
+        return false;
+    }
+
+    fseek(file, 0, SEEK_END); 
+    auto fileSize = ftell(file); 
+    fseek(file, 0, SEEK_SET); 
+
+    auto buffer = (unsigned char *)malloc(fileSize); // Allocate memory to hold the file data
+    if (buffer == NULL) {
+        fprintf(stderr, "%s: failed to alloc %ld bytes for file %s\n", __func__, fileSize, path);
+        perror("Memory allocation error");
+        fclose(file);
+        return false;
+    }
+    fread(buffer, 1, fileSize, file); // Read the file into the buffer
+    fclose(file); // Close the file
+
+    *bytesOut = buffer;
+    *sizeOut = fileSize;
+    return true;
+
+}
+
+LLAMA_API struct llava_image_embed * llava_image_embed_make_with_filename(struct clip_ctx * ctx_clip, int n_threads, const char * image_path)
+{
+    unsigned char* image_bytes;
+    long image_bytes_length;
+    auto loaded = load_file_to_bytes(image_path, &image_bytes, &image_bytes_length);
+    if (!loaded) {
+        fprintf(stderr, "%s: failed to load %s\n", __func__, image_path);
+        return NULL;
+    }
+
+    auto embed = llava_image_embed_make_with_bytes(ctx_clip, n_threads, image_bytes, image_bytes_length);
+    free(image_bytes);
+
+    return embed;
+}
+
+
+LLAMA_API void llava_image_embed_free(struct llava_image_embed * embed) {
+    free(embed->embed);
+    free(embed);
+}