2b-vision/decision.json

{
 "engine": "onnx",
 "family": "decider",
 "layout": "decider-vision-v1",
 "upstream": {
  "repo": "Mapika/decider-2b-vision",
  "revision": "863e290863655f1d6b69324d77d09ac972d21609",
  "code": "decider/vision.py + decider/prompt.py (v5)"
 },
 "graphs": {
  "vision": "vision.onnx",
  "decoder": "model.onnx"
 },
 "max_ctx_tokens": 1536,
 "max_options": 10,
 "max_levels": 10,
 "isolated_levels": false,
 "labels": {
  "narrow": 10,
  "strings": [
   "A",
   "B",
   "C",
   "D",
   "E",
   "F",
   "G",
   "H",
   "I",
   "J"
  ],
  "ids": [
   32,
   33,
   34,
   35,
   36,
   37,
   38,
   39,
   40,
   41
  ],
  "open_ids": [
   198,
   7
  ]
 },
 "chunk": 64,
 "tokens": {
  "image": 248056,
  "vision_start": 248053,
  "vision_end": 248054,
  "pad": 0
 },
 "image": {
  "patch_size": 16,
  "temporal_patch_size": 2,
  "merge_size": 2,
  "min_pixels": 65536,
  "max_pixels": 16777216,
  "resample": "PIL bicubic",
  "rescale_factor": 0.00392156862745098,
  "image_mean": [
   0.5,
   0.5,
   0.5
  ],
  "image_std": [
   0.5,
   0.5,
   0.5
  ],
  "position_table_side": 48,
  "position_interpolation": "bilinear, align_corners"
 },
 "mrope_section": [
  11,
  11,
  10
 ],
 "opset": 20,
 "weights_in_memory": "fp32",
 "min_options": 2,
 "isolated_row_temperature": 1.0,
 "neutralize_none": false,
 "independent_rows": true,
 "index_arrays_min_len": 8,
 "special_tokens": {
  "pad": 0
 },
 "retokenize": "upstream decodes each row's text ids and tokenizes the string again (with the image placeholder in front); rows are encode(decode(ids))"
}