{ "source_model": "/root/code/pulpie-onnx/artifacts/source_model", "output_dir": "/root/code/pulpie-onnx/artifacts/onnx_model", "opset": 17, "sample_input_shape": [ 1, 91 ], "sample_logits_shape": [ 1, 91, 2 ], "sample_logits_dtype": "torch.float32", "raw_model": { "path": "/root/code/pulpie-onnx/artifacts/onnx_model/model_raw.onnx", "checker_passed": true, "error": null, "summary": { "ir_version": 10, "opsets": { "ai.onnx": 18 }, "node_count": 680, "nonstandard_domains": [], "top_ops": { "Mul": 134, "MatMul": 110, "Add": 86, "Transpose": 61, "Reshape": 48, "Slice": 48, "Concat": 28, "Pow": 25, "ReduceMean": 25, "Sqrt": 25, "Reciprocal": 25, "Neg": 24, "Softmax": 12, "Sigmoid": 12, "Unsqueeze": 4, "Cast": 3, "Shape": 2, "Squeeze": 1, "Gather": 1, "Expand": 1, "Sub": 1, "Where": 1, "Range": 1, "Cos": 1, "Sin": 1 } }, "size_bytes": 1463820 }, "optimizer_attempts": [ { "level": "ORT_ENABLE_EXTENDED", "path": "/root/code/pulpie-onnx/artifacts/onnx_model/model_ort_enable_extended.onnx", "standard_onnx": false, "error": "nonstandard domains: ['com.microsoft']", "summary": { "ir_version": 10, "opsets": { "ai.onnx": 18, "ai.onnx.ml": 5, "ai.onnx.training": 1, "ai.onnx.preview.training": 1, "com.microsoft": 1, "com.microsoft.experimental": 1, "com.microsoft.nchwc": 1, "org.pytorch.aten": 1 }, "node_count": 623, "nonstandard_domains": [ "com.microsoft" ], "top_ops": { "Mul": 110, "MatMul": 98, "Add": 86, "Transpose": 52, "Reshape": 48, "Concat": 28, "Pow": 25, "ReduceMean": 25, "Sqrt": 25, "Reciprocal": 25, "Split": 24, "Neg": 24, "com.microsoft::FusedMatMul": 12, "Softmax": 12, "com.microsoft::QuickGelu": 12, "Unsqueeze": 4, "Cast": 3, "Shape": 2, "Squeeze": 1, "Range": 1, "Sin": 1, "Gather": 1, "Cos": 1, "Expand": 1, "Sub": 1, "Where": 1 } } }, { "level": "ORT_ENABLE_BASIC", "path": "/root/code/pulpie-onnx/artifacts/onnx_model/model_ort_enable_basic.onnx", "standard_onnx": true, "error": null, "summary": { "ir_version": 10, "opsets": { "ai.onnx": 18, "ai.onnx.ml": 5, "ai.onnx.training": 1, "ai.onnx.preview.training": 1, "com.microsoft": 1, "com.microsoft.experimental": 1, "com.microsoft.nchwc": 1, "org.pytorch.aten": 1 }, "node_count": 671, "nonstandard_domains": [], "top_ops": { "Mul": 134, "MatMul": 110, "Add": 86, "Transpose": 52, "Reshape": 48, "Slice": 48, "Concat": 28, "Pow": 25, "ReduceMean": 25, "Sqrt": 25, "Reciprocal": 25, "Neg": 24, "Softmax": 12, "Sigmoid": 12, "Unsqueeze": 4, "Cast": 3, "Shape": 2, "Squeeze": 1, "Range": 1, "Sin": 1, "Gather": 1, "Cos": 1, "Expand": 1, "Sub": 1, "Where": 1 } } } ], "model": { "path": "/root/code/pulpie-onnx/artifacts/onnx_model/model.onnx", "checker_passed": true, "summary": { "ir_version": 10, "opsets": { "ai.onnx": 18, "ai.onnx.ml": 5, "ai.onnx.training": 1, "ai.onnx.preview.training": 1, "com.microsoft": 1, "com.microsoft.experimental": 1, "com.microsoft.nchwc": 1, "org.pytorch.aten": 1 }, "node_count": 671, "nonstandard_domains": [], "top_ops": { "Mul": 134, "MatMul": 110, "Add": 86, "Transpose": 52, "Reshape": 48, "Slice": 48, "Concat": 28, "Pow": 25, "ReduceMean": 25, "Sqrt": 25, "Reciprocal": 25, "Neg": 24, "Softmax": 12, "Sigmoid": 12, "Unsqueeze": 4, "Cast": 3, "Shape": 2, "Squeeze": 1, "Range": 1, "Sin": 1, "Gather": 1, "Cos": 1, "Expand": 1, "Sub": 1, "Where": 1 } }, "size_bytes": 847167095 }, "quantization": { "enabled": true, "path": "/root/code/pulpie-onnx/artifacts/onnx_model/model_quantized.onnx", "checker_passed": true, "error": null, "summary": { "ir_version": 10, "opsets": { "ai.onnx": 18, "ai.onnx.ml": 5, "ai.onnx.training": 1, "ai.onnx.preview.training": 1, "com.microsoft": 1, "com.microsoft.experimental": 1, "com.microsoft.nchwc": 1, "org.pytorch.aten": 1 }, "node_count": 976, "nonstandard_domains": [], "top_ops": { "Mul": 304, "Cast": 88, "Add": 86, "MatMulInteger": 85, "Transpose": 52, "DynamicQuantizeLinear": 49, "Reshape": 48, "Slice": 48, "Concat": 28, "Pow": 25, "ReduceMean": 25, "MatMul": 25, "Sqrt": 25, "Reciprocal": 25, "Neg": 24, "Softmax": 12, "Sigmoid": 12, "Unsqueeze": 4, "Shape": 2, "Gather": 1, "Squeeze": 1, "DequantizeLinear": 1, "Range": 1, "Expand": 1, "Sub": 1, "Where": 1, "Sin": 1, "Cos": 1 } }, "size_bytes": 212574154 } }