fxmarty/tiny-gemma-onnx-quantized-trt

1{2  "one_external_file": true,3  "opset": null,4  "optimization": {},5  "optimum_version": "1.18.0.dev0",6  "quantization": {7    "activations_dtype": "QInt8",8    "activations_symmetric": true,9    "format": "QDQ",10    "is_static": true,11    "mode": "QLinearOps",12    "nodes_to_exclude": [],13    "nodes_to_quantize": [],14    "operators_to_quantize": [15      "Conv",16      "ConvTranspose",17      "Gemm",18      "Clip",19      "Relu",20      "Reshape",21      "Transpose",22      "Squeeze",23      "Unsqueeze",24      "Resize",25      "MaxPool",26      "AveragePool",27      "MatMul",28      "Split",29      "Gather",30      "Where",31      "InstanceNormalization",32      "LayerNormalization"33    ],34    "per_channel": false,35    "qdq_add_pair_to_weight": true,36    "qdq_dedicated_pair": true,37    "qdq_op_type_per_channel_support_to_axis": {38      "MatMul": 139    },40    "reduce_range": false,41    "weights_dtype": "QInt8",42    "weights_symmetric": true43  },44  "transformers_version": "4.39.0.dev0",45  "use_external_data_format": true46}47