This is an automated email from the ASF dual-hosted git repository.
masahi pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new 47ef9466b2 [Pytorch] Add quantized::leaky_relu (#11729)
47ef9466b2 is described below
commit 47ef9466b2751acb8c545ea0c3124c70870ec399
Author: yuanfz <[email protected]>
AuthorDate: Thu Jun 16 05:48:52 2022 +0200
[Pytorch] Add quantized::leaky_relu (#11729)
* emptycommit 2nd try
* add operator and test
* example output
* lint with black
* register param index
* remove assert as it is a warning in torch
* fix algo bug
Co-authored-by: yuanfz <[email protected]>
---
python/tvm/relay/frontend/qnn_torch.py | 21 +++++++++++++++++++++
tests/python/frontend/pytorch/qnn_test.py | 14 ++++++++++++++
2 files changed, 35 insertions(+)
diff --git a/python/tvm/relay/frontend/qnn_torch.py
b/python/tvm/relay/frontend/qnn_torch.py
index 41543ec611..63ee6ea96f 100644
--- a/python/tvm/relay/frontend/qnn_torch.py
+++ b/python/tvm/relay/frontend/qnn_torch.py
@@ -271,6 +271,7 @@ def _get_quant_param_for_input(input_value):
"quantized::add_scalar": (2, 3),
"quantized::hardswish": (1, 2),
"quantized::conv_transpose2d": qconv_indices,
+ "quantized::leaky_relu": (3, 4),
}
def dfs(current_node):
@@ -443,6 +444,7 @@ def add_input_quant_params_to_op_inputs(graph):
"quantized::hardswish": 1,
"aten::hardsigmoid": 1,
"quantized::conv_transpose2d": 1,
+ "quantized::leaky_relu": 1,
}
need_input_quant_param = set(num_quantized_inputs.keys())
@@ -935,6 +937,24 @@ def _relu6():
return _impl
+def _leaky_relu():
+ # refer to src/ATen/native/quantized/cpu/qrelu.cpp
+ def _impl(inputs, _):
+ assert len(inputs) == 7, "Input quant params not found in op inputs"
+ alpha = inputs[1]
+ output_scale = _expr.const(inputs[3])
+ output_zero_point = _expr.const(inputs[4])
+ input_scale = _expr.const(inputs[5])
+ input_zero_point = _expr.const(inputs[6])
+ dequant = relay.qnn.op.dequantize(inputs[0], input_scale,
input_zero_point)
+ dequantized = _op.nn.leaky_relu(dequant, alpha)
+ return relay.qnn.op.quantize(
+ dequantized, output_scale, output_zero_point, out_dtype="uint8"
+ )
+
+ return _impl
+
+
def _mul_scalar():
# this is used for mobilenet v3
def _impl(inputs, _):
@@ -1131,6 +1151,7 @@ convert_map = {
"quantized::add_scalar": _add_scalar(),
"quantized::mul_scalar": _mul_scalar(),
"quantized::relu6": _relu6(),
+ "quantized::leaky_relu": _leaky_relu(),
"quantized::linear_dynamic": _linear_dynamic(),
"quantized::hardswish": _hswish(),
"quantized::conv_transpose2d": _quantized_conv_transpose2d(),
diff --git a/tests/python/frontend/pytorch/qnn_test.py
b/tests/python/frontend/pytorch/qnn_test.py
index 6e87b9ee4f..ef7f3f769c 100644
--- a/tests/python/frontend/pytorch/qnn_test.py
+++ b/tests/python/frontend/pytorch/qnn_test.py
@@ -148,6 +148,18 @@ class ReLU(nn.Module):
pass
+class LeakyReLU(nn.Module):
+ def __init__(self):
+ super().__init__()
+ self.leaky_relu = QuantWrapper(nn.LeakyReLU())
+
+ def forward(self, x):
+ return self.leaky_relu(x)
+
+ def fuse_model(self):
+ pass
+
+
# Mobilenet V3 related modules
class Hsigmoid(nn.Module):
def __init__(self, add_stub=False):
@@ -302,6 +314,7 @@ def test_quantized_modules():
("semodule", (1, 16, 64, 64), SqueezeExcite(16, add_stub=True),
False),
("semodule, per_channel", (1, 16, 64, 64), SqueezeExcite(16,
add_stub=True), True),
("mul_scalar negative", imagenet_ishape, MulScalarNegative(),
False),
+ ("leaky_relu", imagenet_ishape, LeakyReLU(), False),
]
for (module_name, ishape, raw_module, per_channel) in qmodules:
@@ -347,6 +360,7 @@ def test_quantized_modules():
# sample outputs
"""
relu 0.0039215684 2.6052087e-08 0.9999933567176871
+ leaky_relu 0.0 0.0 1.0
upsample bilinear 0.0 0.0 1.0
conv_bn 0.22062653 0.011478779 0.6909348115006899
conv_bn_relu 0.3700896 0.010921672 0.7489366477964451
