This is an automated email from the ASF dual-hosted git repository.
junrushao 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 1d39f2c974 [FQ2I] fix unary op output affine type in fq2i (#12224)
1d39f2c974 is described below
commit 1d39f2c974e09e5a767b67e127a5132f0b36c102
Author: Matthew Brookhart <mbrookh...@octoml.ai>
AuthorDate: Sat Jul 30 21:00:55 2022 -0600
[FQ2I] fix unary op output affine type in fq2i (#12224)
* fix unary op output affine type in fq2i
* better names
* add option to force to positive values for ops that are undefined on
negative values
---
.../transform/fake_quantization_to_integer.py | 2 +-
.../test_pass_fake_quantization_to_integer.py | 35 +++++++++++++++-------
2 files changed, 25 insertions(+), 12 deletions(-)
diff --git a/python/tvm/relay/transform/fake_quantization_to_integer.py
b/python/tvm/relay/transform/fake_quantization_to_integer.py
index 8308298e70..b0464439b0 100644
--- a/python/tvm/relay/transform/fake_quantization_to_integer.py
+++ b/python/tvm/relay/transform/fake_quantization_to_integer.py
@@ -534,7 +534,7 @@ def register_unary_qnn(op_name, op):
out_t.scale,
out_t.zero_point,
)
- return [out, x_t]
+ return [out, out_t]
return register_fake_quantization_to_integer(op_name, unary)
diff --git a/tests/python/relay/test_pass_fake_quantization_to_integer.py
b/tests/python/relay/test_pass_fake_quantization_to_integer.py
index d0c8cca6b7..38520ff2df 100644
--- a/tests/python/relay/test_pass_fake_quantization_to_integer.py
+++ b/tests/python/relay/test_pass_fake_quantization_to_integer.py
@@ -318,23 +318,36 @@ def test_fake_quantize_global_avg_pool():
class TestUnaryQNNOp:
- def helper_test_fake_quantize_unary_op(self, fp32_op, scale=0.125):
- x = relay.var("x", shape=[1, 3, 3, 3], dtype="int8")
- mid_point = relay.const(-128)
+ def helper_test_fake_quantize_unary_op(self, fp32_op, pos_values=False):
+ for dtype in ["int8", "uint8"]:
+ x = relay.var("x", shape=[1, 3, 3, 3], dtype=dtype)
- x = relay.qnn.op.dequantize(x, relay.const(scale), mid_point)
- op = fp32_op(x)
- op = relay.qnn.op.quantize(op, relay.const(scale), mid_point)
+ zero = -128 if dtype == "int8" else 0
+ if pos_values:
+ # Use a positive range for quanitzed ops that only work on
positive values
+ input_mid_point = relay.const(zero)
+ output_mid_point = relay.const(zero)
+ else:
+ input_mid_point = relay.const(np.random.randint(0, 255) + zero)
+ output_mid_point = relay.const(np.random.randint(0, 255) +
zero)
- x_np = np.random.randint(-128, 127, size=[1, 3, 3, 3], dtype="int8")
+ input_scale = relay.const(np.random.rand())
+ output_scale = relay.const(np.random.rand())
- compare_fq_to_int(op, [x_np], True)
+ x = relay.qnn.op.dequantize(x, input_scale, input_mid_point)
+ op = fp32_op(x)
+
+ op = relay.qnn.op.quantize(op, output_scale, output_mid_point,
out_dtype=dtype)
+
+ x_np = np.random.randint(0 + zero, 255 + zero, size=[1, 3, 3, 3],
dtype=dtype)
+
+ compare_fq_to_int(op, [x_np], True)
def test_sqrt(self):
- self.helper_test_fake_quantize_unary_op(fp32_op=relay.sqrt)
+ self.helper_test_fake_quantize_unary_op(fp32_op=relay.sqrt,
pos_values=True)
def test_rsqrt(self):
- self.helper_test_fake_quantize_unary_op(fp32_op=relay.rsqrt)
+ self.helper_test_fake_quantize_unary_op(fp32_op=relay.rsqrt,
pos_values=True)
def test_exp(self):
self.helper_test_fake_quantize_unary_op(fp32_op=relay.exp)
@@ -349,7 +362,7 @@ class TestUnaryQNNOp:
self.helper_test_fake_quantize_unary_op(fp32_op=relay.tanh)
def test_log(self):
- self.helper_test_fake_quantize_unary_op(fp32_op=relay.log)
+ self.helper_test_fake_quantize_unary_op(fp32_op=relay.log,
pos_values=True)
def test_fake_quantize_reshape():
