Downloads · 30 days
12
24% of all-time downloads
axiom-of-choice/OmniParser-v2-icon-caption-4bit
OmniParser-v2-icon-caption-4bit is a machine learning model from axiom-of-choice. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for mlx. The card lists the license as mit.
MLX conversion of the icon captioning model from microsoft/OmniParser-v2.0.
Downloads · 30 days
12
24% of all-time downloads
All-time downloads
50
Public
Parameters
271M
284 MB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors284 MB · 98%
How the weights are stored.
U32179M · 66%
From the Hugging Face model README
MLX conversion of the icon captioning model from microsoft/OmniParser-v2.0.
from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
model, processor = load("axiom-of-choice/OmniParser-v2-icon-caption-4-bit (group_size=64)")
prompt = apply_chat_template(processor, model.config, "<CAPTION>", num_images=1)
output = generate(model, processor, prompt, image=crop, max_tokens=20, temperature=0.0)
print(output.text)
Fine-tuned weights from microsoft/OmniParser-v2.0/icon_caption, processor from microsoft/Florence-2-base.
Config patched for transformers 5.x compatibility (auto_map removed, text_config.model_type set to "bart").