mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2025-10-31 08:51:55 +00:00
7084755396
This is a follup of Commit fc0c8d286a
("llava : update surgery script to not remove tensors") but this time
the change is to the BakLLaVA specific part of the surgery script.
I've been able to test this using SkunkworksAI/BakLLaVA-1 and it works
as expected using the instructions in README.md.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
39 lines
1.3 KiB
Python
39 lines
1.3 KiB
Python
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.")
|