{ "image_size": 640, "crop_mode": false, "patch_grid": 10, "visual_tokens": 111, "prefix_len": 115, "image_token_id": 128815, "bos": 0, "eos": 1, "hidden": 1280, "preprocess": "ImageOps.pad to 640x640 (pad color = mean*255), normalize mean=std=0.5", "arrange": "vision[100,1280] -> view[10,10] -> +image_newline per row [110] -> +view_seperator [111]", "scatter": "embed_tokens(prompt_input_ids); fill input_ids==128815 positions (in order) with the 111 visual features", "files": {"embed": "embed_tokens.f16 [129280,1280] fp16", "newline": "image_newline.f16 [1280]", "seperator": "view_seperator.f16 [1280]", "prompt": "prompt_input_ids.i32 [115]"} }