Merge mainline llama.cpp (#3)

* Merging mainline - WIP

* Merging mainline - WIP

AVX2 and CUDA appear to work.
CUDA performance seems slightly (~1-2%) lower as it is so often
the case with llama.cpp/ggml after some "improvements" have been made.

* Merging mainline - fix Metal

* Remove check

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>

1 files changed, 0 insertions, 38 deletions

diff --git a/examples/llava/llava-surgery.py b/examples/llava/llava-surgery.py
deleted file mode 100644
index 4f2da3be..00000000
--- a/examples/llava/llava-surgery.py
+++ /dev/null
@@ -1,38 +0,0 @@
-import argparse
-import glob
-import os
-import torch
-
-
-ap = argparse.ArgumentParser()
-ap.add_argument("-m", "--model", help="Path to LLaVA v1.5 model")
-args = ap.parse_args()
-
-# find the model part that includes the the multimodal projector weights
-path = sorted(glob.glob(f"{args.model}/pytorch_model*.bin"))[-1]
-checkpoint = torch.load(path)
-
-# get a list of mm tensor names
-mm_tensors = [k for k, v in checkpoint.items() if k.startswith("model.mm_projector")]
-
-# store these tensors in a new dictionary and torch.save them
-projector = {name: checkpoint[name].float() for name in mm_tensors}
-torch.save(projector, f"{args.model}/llava.projector")
-
-# BakLLaVA models contain CLIP tensors in it
-clip_tensors = [k for k, v in checkpoint.items() if k.startswith("model.vision_tower")]
-if len(clip_tensors) > 0:
-    clip = {name.replace("vision_tower.vision_tower.", ""): checkpoint[name].float() for name in clip_tensors}
-    torch.save(clip, f"{args.model}/llava.clip")
-
-
-    # added tokens should be removed to be able to convert Mistral models
-    if os.path.exists(f"{args.model}/added_tokens.json"):
-        with open(f"{args.model}/added_tokens.json", "w") as f:
-            f.write("{}\n")
-
-
-
-print("Done!")
-print(f"Now you can convert {args.model} to a regular LLaMA GGUF file.")
-print(f"Also, use {args.model}/llava.projector to prepare a llava-encoder.gguf file.")