- Sort Score
- Result 10 results
- Languages All
Results 1 - 8 of 8 for numBits (0.11 sec)
-
tensorflow/compiler/mlir/quantization/common/ir/FakeQuantSupport.cc
namespace mlir::quantfork { static bool getDefaultStorageParams(unsigned numBits, bool narrowRange, bool isSigned, MLIRContext *ctx, Type &storageType, int64_t &qmin, int64_t &qmax) { // Hard-coded type mapping from TFLite. if (numBits <= 4) { storageType = IntegerType::get(ctx, 4); if (isSigned) { qmin = -8;
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Mar 21 11:52:27 UTC 2024 - 7.7K bytes - Viewed (0) -
tensorflow/compiler/mlir/quantization/common/ir/FakeQuantSupport.h
// // Specifically, it combines the following concerns, each of which would be // independent variables in a more generic setup: // - numBits and isSigned imply storage data type (uint8, int8, int16) // - numBits < 8 is promoted to uint8 or int8 // - "narrow_range" narrows the lower bound of the storage type's range by // 1
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Mar 21 11:52:27 UTC 2024 - 3.7K bytes - Viewed (0) -
tensorflow/compiler/mlir/lite/experimental/tac/transforms/device_transform.cc
const int64_t int_value = ap_int_value.getSExtValue(); const float real = (int_value - zp) * scale; auto real_int = absl::bit_cast<int32_t>(real); return APInt(/*numBits=*/32, real_int); }; auto dequant_values = mlir::cast<DenseIntOrFPElementsAttr>(input_values) .mapValues(FloatType::getF32(rewriter.getContext()),
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Apr 25 16:01:03 UTC 2024 - 8.4K bytes - Viewed (0) -
tensorflow/compiler/mlir/lite/transforms/default_quant_params.cc
quant::QuantParams DefaultQuantParamsPass::GetDefaultQuantParams( Builder builder) { if (!default_quant_params_) { default_quant_params_ = quantfork::fakeQuantAttrsToType( builder.getUnknownLoc(), /*numBits=*/8, default_min_, default_max_, /*narrowRange=*/false, builder.getF32Type(), is_signed_); } return default_quant_params_; } // Creates an instance of the default quant parameters pass.
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Apr 25 16:01:03 UTC 2024 - 9.4K bytes - Viewed (0) -
tensorflow/compiler/mlir/lite/tests/mlir2flatbuffer/fake_quant.mlir
// CHECK-NEXT: } ] // CHECK-NEXT: signature_defs: [ ] // CHECK-NEXT: } // IMPORT: "tfl.fake_quant"(%arg0) <{max = 1.400000e+00 : f32, min = 3.000000e-01 : f32, narrow_range = false, num_bits = 6 : i32}> %0 = "tfl.fake_quant"(%arg0) {num_bits = 6 : i32, narrow_range = false, min = 0.3:f32, max = 1.4:f32} : (tensor<4 x f32>) -> tensor<4 x f32> func.return %0 : tensor<4xf32>
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu May 02 09:41:17 UTC 2024 - 2.4K bytes - Viewed (0) -
tensorflow/compiler/mlir/lite/utils/fake_quant_utils.h
quant_dim = mlir::cast<ShapedType>(res.getType()).getRank() - 1; } // Use the min/max from the operands and the num_bits and narrow_range // attribute to create the quantization parameter for the new quantize op. rewriter.setInsertionPointAfter(tf_op.getOperation()); IntegerAttr num_bits = rewriter.getI64IntegerAttr(tf_op.getNumBits()); BoolAttr narrow_range = rewriter.getBoolAttr(tf_op.getNarrowRange());
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Apr 25 16:01:03 UTC 2024 - 6.6K bytes - Viewed (0) -
tensorflow/compiler/mlir/lite/quantization/tensorflow/tf_to_quant.cc
quant_dim = mlir::cast<ShapedType>(res.getType()).getRank() - 1; } // Use the min/max from the operands and the num_bits and narrow_range // attribute to create the quantization parameter for the new quantize op. rewriter.setInsertionPointAfter(tf_op.getOperation()); IntegerAttr num_bits = rewriter.getI64IntegerAttr(tf_op.getNumBits()); BoolAttr narrow_range = rewriter.getBoolAttr(tf_op.getNarrowRange());
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Apr 25 16:01:03 UTC 2024 - 8.1K bytes - Viewed (0) -
tensorflow/compiler/mlir/quantization/tensorflow/utils/fake_quant_utils.h
quant_dim = input_type.getRank() - 1; } // Use the min/max from the operands and the num_bits and narrow_range // attribute to create the quantization parameter for the new quantize op. rewriter.setInsertionPointAfter(tf_op.getOperation()); IntegerAttr num_bits = rewriter.getI64IntegerAttr(tf_op.getNumBits()); BoolAttr narrow_range = rewriter.getBoolAttr(tf_op.getNarrowRange());
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Apr 25 16:01:03 UTC 2024 - 6.3K bytes - Viewed (0)