OS-Copilot/OS-Atlas-Pro-4B https://huggingface.co/OS-Copilot/OS-Atlas-Pro-4B
OS-Copilot/OS-Atlas-Pro-4B https://huggingface.co/OS-Copilot/OS-Atlas-Pro-4B
There is already mradermacher/OS-Atlas-Pro-7B-GGUF, but OS-Atlas-Pro-4B is lighter and can run on many PCs.
It's queued! :D
Looks like a realy interesting model.
You can check for progress at http://hf.tst.eu/status.html or regularly check the model
summary page at https://hf.tst.eu/model#OS-Atlas-Pro-4B-GGUF for quants to appear.
The for the mmproj extraction required preprocessor_config.json file is missing from the model. I decided to use https://huggingface.co/OpenGVLab/InternVL2-4B/resolve/main/preprocessor_config.json with the following content instead as I'm like 99% sure that they would have just copied it anyways:
{
"crop_size": 448,
"do_center_crop": true,
"do_normalize": true,
"do_resize": true,
"feature_extractor_type": "CLIPFeatureExtractor",
"image_mean": [
0.485,
0.456,
0.406
],
"image_std": [
0.229,
0.224,
0.225
],
"resample": 3,
"size": 448
}
We unfortunately just got another error and this time it is proxbably not fixable: ValueError: Can not map tensor 'language_model.model.layers.0.input_layernorm.weight'