graph_report.json

{
  "package": "esm2-t6-8m",
  "providers": [
    "AzureExecutionProvider",
    "CPUExecutionProvider"
  ],
  "graphs": [
    {
      "path": "model.onnx",
      "size": 94903,
      "nodes": 142,
      "ops": {
        "ai.onnx::Add": 36,
        "ai.onnx::Attention": 6,
        "ai.onnx::Cast": 3,
        "ai.onnx::Concat": 2,
        "ai.onnx::Constant": 3,
        "ai.onnx::Div": 2,
        "ai.onnx::Equal": 1,
        "ai.onnx::Expand": 2,
        "ai.onnx::Gather": 3,
        "ai.onnx::Gelu": 6,
        "ai.onnx::LayerNormalization": 1,
        "ai.onnx::MatMul": 36,
        "ai.onnx::Mul": 2,
        "ai.onnx::Range": 1,
        "ai.onnx::ReduceSum": 2,
        "ai.onnx::RotaryEmbedding": 12,
        "ai.onnx::Shape": 3,
        "ai.onnx::Squeeze": 1,
        "ai.onnx::Sub": 1,
        "ai.onnx::Unsqueeze": 6,
        "ai.onnx::Where": 1,
        "com.microsoft::SkipLayerNormalization": 12
      },
      "inputs": [
        {
          "name": "input_ids",
          "dtype": "INT64",
          "shape": "[component.model.batch,component.model.sequence_len]"
        },
        {
          "name": "attention_mask",
          "dtype": "INT64",
          "shape": "[component.model.batch,component.model.sequence_len]"
        }
      ],
      "outputs": [
        {
          "name": "last_hidden_state",
          "dtype": "FLOAT",
          "shape": "[component.model.batch,component.model.sequence_len,320]"
        }
      ],
      "initializers": 108,
      "initializer_bytes": 29701556,
      "ort_load": {
        "ok": true,
        "seconds": 0.1396461280528456,
        "providers": [
          "CPUExecutionProvider"
        ]
      }
    }
  ]
}