workflow.json

{
  "schema_version": "2",
  "name": "North Micro Vision Instruct",
  "description": "Visual question answering, captioning, OCR, and grounding with CohereLabs/North-Micro-Vision-Instruct (2.4B native-resolution VLM) on ZeroGPU, driven by a gr.Workflow fn-bound @spaces.GPU function. Single Space: the workflow frontend and the GPU worker share one process.",
  "runtime": { "default": "client" },
  "view": { "default": "canvas" },
  "references": [
    {
      "id": "ref_image",
      "label": "Input Image",
      "role": "reference",
      "asset_type": "image",
      "inputs": [{ "id": "in", "label": "Image", "type": "image" }],
      "outputs": [{ "id": "out", "label": "Image", "type": "image" }],
      "x": 60,
      "y": 120,
      "width": 240,
      "height": 140,
      "data": {}
    },
    {
      "id": "ref_prompt",
      "label": "Prompt",
      "role": "reference",
      "asset_type": "text",
      "inputs": [{ "id": "in", "label": "Prompt", "type": "text" }],
      "outputs": [{ "id": "out", "label": "Prompt", "type": "text" }],
      "x": 60,
      "y": 300,
      "width": 240,
      "height": 120,
      "data": { "out": "What do you see in this image?" }
    },
    {
      "id": "ref_max_tokens",
      "label": "Max New Tokens",
      "role": "reference",
      "asset_type": "number",
      "inputs": [{ "id": "in", "label": "Max New Tokens", "type": "number" }],
      "outputs": [{ "id": "out", "label": "Max New Tokens", "type": "number" }],
      "x": 60,
      "y": 460,
      "width": 200,
      "height": 90,
      "data": { "out": 256 }
    },
    {
      "id": "ref_temperature",
      "label": "Temperature",
      "role": "reference",
      "asset_type": "number",
      "inputs": [{ "id": "in", "label": "Temperature", "type": "number" }],
      "outputs": [{ "id": "out", "label": "Temperature", "type": "number" }],
      "x": 60,
      "y": 570,
      "width": 200,
      "height": 90,
      "data": { "out": 0.7 }
    },
    {
      "id": "ref_top_p",
      "label": "Top P",
      "role": "reference",
      "asset_type": "number",
      "inputs": [{ "id": "in", "label": "Top P", "type": "number" }],
      "outputs": [{ "id": "out", "label": "Top P", "type": "number" }],
      "x": 60,
      "y": 680,
      "width": 200,
      "height": 90,
      "data": { "out": 0.8 }
    },
    {
      "id": "ref_top_k",
      "label": "Top K",
      "role": "reference",
      "asset_type": "number",
      "inputs": [{ "id": "in", "label": "Top K", "type": "number" }],
      "outputs": [{ "id": "out", "label": "Top K", "type": "number" }],
      "x": 60,
      "y": 790,
      "width": 200,
      "height": 90,
      "data": { "out": 20 }
    }
  ],
  "operators": [
    {
      "id": "op_vlm",
      "label": "run_vlm",
      "role": "operator",
      "kind": "fn",
      "source": "fn",
      "fn": "run_vlm",
      "inputs": [
        { "id": "in_0", "label": "image", "type": "image", "required": true },
        { "id": "in_1", "label": "prompt", "type": "text", "required": true },
        { "id": "in_2", "label": "max_new_tokens", "type": "number" },
        { "id": "in_3", "label": "temperature", "type": "number" },
        { "id": "in_4", "label": "top_p", "type": "number" },
        { "id": "in_5", "label": "top_k", "type": "number" }
      ],
      "outputs": [
        { "id": "out_0", "label": "response", "type": "text", "output_index": 0 }
      ],
      "x": 420,
      "y": 380,
      "width": 280,
      "height": 220,
      "data": {}
    }
  ],
  "subjects": [
    {
      "id": "sub_response",
      "label": "Response",
      "role": "subject",
      "asset_type": "text",
      "inputs": [{ "id": "in", "label": "Response", "type": "text" }],
      "outputs": [{ "id": "out", "label": "Response", "type": "text" }],
      "x": 820,
      "y": 380,
      "width": 280,
      "height": 140,
      "data": {}
    }
  ],
  "edges": [
    { "id": "e_image",       "from_node_id": "ref_image",       "from_port_id": "out",   "to_node_id": "op_vlm",       "to_port_id": "in_0", "type": "image" },
    { "id": "e_prompt",      "from_node_id": "ref_prompt",      "from_port_id": "out",   "to_node_id": "op_vlm",       "to_port_id": "in_1", "type": "text" },
    { "id": "e_max_tokens",  "from_node_id": "ref_max_tokens",  "from_port_id": "out",   "to_node_id": "op_vlm",       "to_port_id": "in_2", "type": "number" },
    { "id": "e_temperature", "from_node_id": "ref_temperature", "from_port_id": "out",   "to_node_id": "op_vlm",       "to_port_id": "in_3", "type": "number" },
    { "id": "e_top_p",       "from_node_id": "ref_top_p",       "from_port_id": "out",   "to_node_id": "op_vlm",       "to_port_id": "in_4", "type": "number" },
    { "id": "e_top_k",       "from_node_id": "ref_top_k",       "from_port_id": "out",   "to_node_id": "op_vlm",       "to_port_id": "in_5", "type": "number" },
    { "id": "e_response",    "from_node_id": "op_vlm",          "from_port_id": "out_0", "to_node_id": "sub_response", "to_port_id": "in",   "type": "text" }
  ]
}