From cc6cac08e38e32bf40bbe07e9e8f8f0130b5fd94 Mon Sep 17 00:00:00 2001 From: Daniel Bevenius Date: Wed, 21 Feb 2024 14:36:57 +0100 Subject: [PATCH] llava : add --skip-unknown to 1.6 convert.py (#5632) This commit adds the `--skip-unknown` option to the convert.py script and removes the saving of the updated checkpoints to avoid updating possibly checked out files. The motivation for this change is that this was done for 1.5 in Commit fc0c8d286a533363a9a663510b62af85ffad58b3 ("llava : update surgery script to not remove tensors") and makes the examples more consistent. Signed-off-by: Daniel Bevenius --- examples/llava/README.md | 13 ++++++------- examples/llava/llava-surgery-v2.py | 12 ------------ 2 files changed, 6 insertions(+), 19 deletions(-) diff --git a/examples/llava/README.md b/examples/llava/README.md index 25ea96715..35e6d9e5d 100644 --- a/examples/llava/README.md +++ b/examples/llava/README.md @@ -63,13 +63,12 @@ Now both the LLaMA part and the image encoder is in the `llava-v1.5-7b` director ```console git clone https://huggingface.co/liuhaotian/llava-v1.6-vicuna-7b ``` -2) Backup your pth/safetensor model files as llava-surgery modifies them -3) Use `llava-surgery-v2.py` which also supports llava-1.5 variants pytorch as well as safetensor models: +2) Use `llava-surgery-v2.py` which also supports llava-1.5 variants pytorch as well as safetensor models: ```console python examples/llava/llava-surgery-v2.py -C -m ../llava-v1.6-vicuna-7b/ ``` - you will find a llava.projector and a llava.clip file in your model directory -4) Copy the llava.clip file into a subdirectory (like vit), rename it to pytorch_model.bin and add a fitting vit configuration to the directory: +3) Copy the llava.clip file into a subdirectory (like vit), rename it to pytorch_model.bin and add a fitting vit configuration to the directory: ```console mkdir vit cp ../llava-v1.6-vicuna-7b/llava.clip vit/pytorch_model.bin @@ -77,18 +76,18 @@ cp ../llava-v1.6-vicuna-7b/llava.projector vit/ curl -s -q https://huggingface.co/cmp-nct/llava-1.6-gguf/raw/main/config_vit.json -o vit/config.json ``` -5) Create the visual gguf model: +4) Create the visual gguf model: ```console python ./examples/llava/convert-image-encoder-to-gguf.py -m vit --llava-projector vit/llava.projector --output-dir vit --clip-model-is-vision ``` - This is similar to llava-1.5, the difference is that we tell the encoder that we are working with the pure vision model part of CLIP -6) Then convert the model to gguf format: +5) Then convert the model to gguf format: ```console -python ./convert.py ../llava-v1.6-vicuna-7b/ +python ./convert.py ../llava-v1.6-vicuna-7b/ --skip-unknown ``` -7) And finally we can run the llava-cli using the 1.6 model version: +6) And finally we can run the llava-cli using the 1.6 model version: ```console ./llava-cli -m ../llava-v1.6-vicuna-7b/ggml-model-f16.gguf --mmproj vit/mmproj-model-f16.gguf --image some-image.jpg -c 4096 ``` diff --git a/examples/llava/llava-surgery-v2.py b/examples/llava/llava-surgery-v2.py index 5bc5bc513..eb56d6988 100644 --- a/examples/llava/llava-surgery-v2.py +++ b/examples/llava/llava-surgery-v2.py @@ -65,9 +65,7 @@ def clean_vision_tower_from_checkpoint(checkpoint_path): for name in clip_tensors: del checkpoint[name] - # Save the updated checkpoint checkpoint_path = checkpoint_path - save_model(checkpoint, checkpoint_path, file_type) return True return False @@ -152,16 +150,6 @@ for name in first_mm_tensors: if len(projector) > 0: save_model(projector, f"{args.model}/llava.projector", 'pytorch') -for name in mm_tensors: - del last_checkpoint[name] -for name in first_mm_tensors: - del first_checkpoint[name] - -if len(mm_tensors) > 0: - save_model(last_checkpoint, projector_checkpoint_path, file_type) -if len(first_mm_tensors) > 0: - save_model(first_checkpoint, newline_checkpoint_path, file_type) - print("Done!") print(f"Now you can convert {args.model} to a a regular LLaMA GGUF file.") print(f"Also, use {args.model}/llava.projector to prepare a llava-encoder.gguf file.")