examples: support LLaVA v1.5 (multimodal model) (#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 <ggerganov@gmail.com>

1 files changed, 73 insertions, 0 deletions

diff --git a/examples/llava/clip.h b/examples/llava/clip.h
new file mode 100644
index 00000000..3d7261e2
--- /dev/null
+++ b/examples/llava/clip.h
@@ -0,0 +1,73 @@
+#ifndef CLIP_H
+#define CLIP_H
+
+#include "ggml.h"
+
+struct clip_ctx;
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct clip_vision_hparams {
+    int32_t image_size;
+    int32_t patch_size;
+    int32_t hidden_size;
+    int32_t n_intermediate;
+    int32_t projection_dim;
+    int32_t n_head;
+    int32_t n_layer;
+    float eps;
+};
+
+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);
+
+// RGB uint8 image
+struct clip_image_u8 {
+    int nx;
+    int ny;
+    uint8_t * data;
+    size_t size;
+};
+
+// RGB float32 image (NHWC)
+// Memory layout: RGBRGBRGB...
+struct clip_image_f32 {
+    int nx;
+    int ny;
+    float * data;
+    size_t size;
+};
+
+struct clip_image_u8_batch {
+    struct clip_image_u8 * data;
+    size_t size;
+};
+
+struct clip_image_f32_batch {
+    struct clip_image_f32 * data;
+    size_t size;
+};
+
+struct clip_image_u8 * make_clip_image_u8();
+struct clip_image_f32 * make_clip_image_f32();
+bool clip_image_load_from_file(const char * fname, 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);
+
+bool clip_image_batch_encode(const struct clip_ctx * ctx, const int n_threads, const struct clip_image_f32_batch * imgs,
+                             float * vec);
+
+bool clip_model_quantize(const char * fname_inp, const char * fname_out, const int itype);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif // CLIP_H