You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2021/12/06 09:39:19 UTC
[GitHub] [tvm] KennyTang1988 opened a new pull request #9655: [Caffe Frontend] Add support for Power layer
KennyTang1988 opened a new pull request #9655:
URL: https://github.com/apache/tvm/pull/9655
Add support for Power layer
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] mshr-h commented on pull request #9655: [Caffe Frontend] Add support for Power layer
Posted by GitBox <gi...@apache.org>.
mshr-h commented on pull request #9655:
URL: https://github.com/apache/tvm/pull/9655#issuecomment-1007209093
@fernchen @FrozenGene
Can you take a look at this PR, please? Thanks!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] KennyTang1988 commented on pull request #9655: [Caffe Frontend] Add support for Power layer
Posted by GitBox <gi...@apache.org>.
KennyTang1988 commented on pull request #9655:
URL: https://github.com/apache/tvm/pull/9655#issuecomment-1023995316
> @fernchen @FrozenGene Can you take a look at this PR, please? Thanks!
I hava update commit, please help check it again.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] masahi merged pull request #9655: [Caffe Frontend] Add support for Power layer
Posted by GitBox <gi...@apache.org>.
masahi merged pull request #9655:
URL: https://github.com/apache/tvm/pull/9655
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] mshr-h commented on a change in pull request #9655: [Caffe Frontend] Add support for Power layer
Posted by GitBox <gi...@apache.org>.
mshr-h commented on a change in pull request #9655:
URL: https://github.com/apache/tvm/pull/9655#discussion_r780078251
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
+ if power != 1.0:
+ out = _op.power(out, power)
Review comment:
Same here.
##########
File path: tests/python/frontend/caffe/test_forward.py
##########
@@ -613,6 +613,24 @@ def test_forward_Pooling():
_test_pooling(data, pool=P.Pooling.AVE, global_pooling=True)
+#######################################################################
+# Power
+# -----------
+def _test_power(data, **kwargs):
+ """One iteration of Power."""
+ _test_op(data, L.Power, "Power", **kwargs)
+
+
+def test_forward_Power():
+ """Power"""
+ data = np.random.rand(1, 3, 10, 10).astype(np.float32)
+ _test_power(data, power_param={"power": 0.37, "scale": 0.83, "shift": -2.4})
+ _test_power(data, power_param={"power": 0.37, "scale": 0.83, "shift": 0.0})
+ _test_power(data, power_param={"power": 0.0, "scale": 0.83, "shift": -2.4})
+ _test_power(data, power_param={"power": 1.0, "scale": 0.83, "shift": -2.4})
+ _test_power(data, power_param={"power": 2.0, "scale": 0.34, "shift": -2.4})
+
Review comment:
Might be better to add a test with Caffe's default parameters.
```python
_test_power(data, power_param={"power": 1.0, "scale": 1.0, "shift": 0.0})
```
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
+ if power != 1.0:
+ out = _op.power(out, power)
+ return out
Review comment:
I think we don't need to check the parameters.
Just like below.
```python
out = _op.multiply(in_expr, scale)
out = _op.add(out, shift)
out = _op.power(out, power)
return out
```
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
Review comment:
if `scale` is 1.0, line 650 won't be executed.
Then `out` is undefined which causes an undefined reference error.
##########
File path: tests/python/frontend/caffe/test_forward.py
##########
@@ -613,6 +613,24 @@ def test_forward_Pooling():
_test_pooling(data, pool=P.Pooling.AVE, global_pooling=True)
+#######################################################################
+# Power
+# -----------
Review comment:
Can you remove unnecessary redundant `--` and keep aligned with Power?
```python
#######################################################################
# Power
# -----
```
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
+ if power != 1.0:
+ out = _op.power(out, power)
+ return out
Review comment:
If `scale` is 1.0 and `shift` is 0.0 and `power` is 1.0, `out` will be undefined.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] mshr-h commented on pull request #9655: [Caffe Frontend] Add support for Power layer
Posted by GitBox <gi...@apache.org>.
mshr-h commented on pull request #9655:
URL: https://github.com/apache/tvm/pull/9655#issuecomment-1007209093
@fernchen @FrozenGene
Can you take a look at this PR, please? Thanks!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] mshr-h commented on a change in pull request #9655: [Caffe Frontend] Add support for Power layer
Posted by GitBox <gi...@apache.org>.
mshr-h commented on a change in pull request #9655:
URL: https://github.com/apache/tvm/pull/9655#discussion_r780078251
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
+ if power != 1.0:
+ out = _op.power(out, power)
Review comment:
Same here.
##########
File path: tests/python/frontend/caffe/test_forward.py
##########
@@ -613,6 +613,24 @@ def test_forward_Pooling():
_test_pooling(data, pool=P.Pooling.AVE, global_pooling=True)
+#######################################################################
+# Power
+# -----------
+def _test_power(data, **kwargs):
+ """One iteration of Power."""
+ _test_op(data, L.Power, "Power", **kwargs)
+
+
+def test_forward_Power():
+ """Power"""
+ data = np.random.rand(1, 3, 10, 10).astype(np.float32)
+ _test_power(data, power_param={"power": 0.37, "scale": 0.83, "shift": -2.4})
+ _test_power(data, power_param={"power": 0.37, "scale": 0.83, "shift": 0.0})
+ _test_power(data, power_param={"power": 0.0, "scale": 0.83, "shift": -2.4})
+ _test_power(data, power_param={"power": 1.0, "scale": 0.83, "shift": -2.4})
+ _test_power(data, power_param={"power": 2.0, "scale": 0.34, "shift": -2.4})
+
Review comment:
Might be better to add a test with Caffe's default parameters.
```python
_test_power(data, power_param={"power": 1.0, "scale": 1.0, "shift": 0.0})
```
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
+ if power != 1.0:
+ out = _op.power(out, power)
+ return out
Review comment:
I think we don't need to check the parameters.
Just like below.
```python
out = _op.multiply(in_expr, scale)
out = _op.add(out, shift)
out = _op.power(out, power)
return out
```
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
Review comment:
if `scale` is 1.0, line 650 won't be executed.
Then `out` is undefined which causes an undefined reference error.
##########
File path: tests/python/frontend/caffe/test_forward.py
##########
@@ -613,6 +613,24 @@ def test_forward_Pooling():
_test_pooling(data, pool=P.Pooling.AVE, global_pooling=True)
+#######################################################################
+# Power
+# -----------
Review comment:
Can you remove unnecessary redundant `--` and keep aligned with Power?
```python
#######################################################################
# Power
# -----
```
##########
File path: python/tvm/relay/frontend/caffe.py
##########
@@ -637,6 +638,22 @@ def convert_embed(self, op):
return out
+ def convert_power(self, op):
+ """Convert Power layer"""
+ inputs = op.bottom
+ in_expr = self.exp_tab.get_expr(inputs[0])
+ power = _expr.const(op.power_param.power)
+ scale = _expr.const(op.power_param.scale)
+ shift = _expr.const(op.power_param.shift)
+
+ if scale != 1.0:
+ out = _op.multiply(in_expr, scale)
+ if shift != 0.0:
+ out = _op.add(out, shift)
+ if power != 1.0:
+ out = _op.power(out, power)
+ return out
Review comment:
If `scale` is 1.0 and `shift` is 0.0 and `power` is 1.0, `out` will be undefined.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org