examples: support LLaVA v1.5 (multimodal model) (ggerganov#3436)

* WIP: start implementing LLaVA

* rm scratch buf for now, will revert after cleanup

* LLaVA image encoder is working. will combine with llama

* Add llava inference code, but it's buggy. debugging

* LLaVA is working e2e, needs to optimize memory allocation + cleanup

* Use ggml_allocr + rm unnecessary code

* fix: crlf -> lf

* fix: new line at EoF

* fix: trailing whitespace

* Add readme

* Update readme

* Some cleanup

* Are you happy editorconfig?

* rm unused batch image preprocessing

* rm unused import

* fix: rm designated initializers

* introduce pad-to-square mode for non-square images

* are you happy editorconfig?

* gitignore /llava

* Handle cases where image file does not exist

* add llava target to Makefile

* add support for 13b model variant

* Maybe seed is unlucky?

* Check if apples are compared to apples

* are you happy editorconfig?

* Use temperature = 0.1 by default

* command line: use gpt_params_parse()

* minor

* handle default n_predict

* fix typo

* llava : code formatting, rename files, fix compile warnings

* do not use Wno-cast-qual for MSVC

---------

Co-authored-by: Georgi Gerganov <[email protected]>

.gitignore

@@ -44,6 +44,7 @@ models-mnt
 /infill
 /libllama.so
 /llama-bench
+/llava
 /main
 /metal
 /perplexity

Makefile

@@ -1,7 +1,7 @@
 # Define the default target now so that it is always the first target
 BUILD_TARGETS = \
 	main quantize quantize-stats perplexity embedding vdot q8dot train-text-from-scratch convert-llama2c-to-ggml \
-	simple batched batched-bench save-load-state server embd-input-test gguf llama-bench baby-llama beam-search  \
+	simple batched batched-bench save-load-state server embd-input-test gguf llama-bench llava baby-llama beam-search  \
 	speculative infill benchmark-matmult parallel finetune export-lora tests/test-c.o
 
 # Binaries only useful for tests
@@ -627,6 +627,9 @@ convert-llama2c-to-ggml: examples/convert-llama2c-to-ggml/convert-llama2c-to-ggm
 llama-bench: examples/llama-bench/llama-bench.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
+llava: examples/llava/llava.cpp examples/llava/llava-utils.h examples/llava/clip.cpp examples/llava/clip.h common/stb_image.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
+	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS) -Wno-cast-qual
+
 baby-llama: examples/baby-llama/baby-llama.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 

common/common.cpp

@@ -384,6 +384,18 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
                 break;
             }
             params.lora_base = argv[i];
+        } else if (arg == "--mmproj") {
+            if (++i >= argc) {
+                invalid_param = true;
+                break;
+            }
+            params.mmproj = argv[i];
+        } else if (arg == "--image") {
+            if (++i >= argc) {
+                invalid_param = true;
+                break;
+            }
+            params.image = argv[i];
         } else if (arg == "-i" || arg == "--interactive") {
             params.interactive = true;
         } else if (arg == "--embedding") {
@@ -703,6 +715,8 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
     printf("  -np N, --parallel N   number of parallel sequences to decode (default: %d)\n", params.n_parallel);
     printf("  -ns N, --sequences N  number of sequences to decode (default: %d)\n", params.n_sequences);
     printf("  -cb, --cont-batching  enable continuous batching (a.k.a dynamic batching) (default: disabled)\n");
+    printf("  --mmproj MMPROJ_FILE  path to a multimodal projector file for LLaVA. see examples/llava/README.md\n");
+    printf("  --image IMAGE_FILE    path to an image file. use with multimodal models\n");
     if (llama_mlock_supported()) {
         printf("  --mlock               force system to keep model in RAM rather than swapping or compressing\n");
     }

common/common.h

@@ -104,6 +104,10 @@ struct gpt_params {
     bool numa              = false; // attempt optimizations that help on some NUMA systems
     bool verbose_prompt    = false; // print prompt tokens before generation
     bool infill            = false; // use infill mode
+
+    // multimodal models (see examples/llava)
+    std::string mmproj = ""; // path to multimodal projector
+    std::string image = ""; // path to an image file
 };
 
 bool gpt_params_parse(int argc, char ** argv, gpt_params & params);