Merge pull request #216 from Xilinx/feature/brevitas_finn_codebug

Brevitas-FINN co-debug

docker/finn_entrypoint.sh

@@ -12,7 +12,7 @@ gecho () {
 
 # checkout the correct dependency repo commits
 # the repos themselves are cloned in the Dockerfile
-BREVITAS_COMMIT=f9a27226d4acf1661dd38bc449f71f89e0983cce
+BREVITAS_COMMIT=172e423164402a07826877fa9730063bee10a208
 CNPY_COMMIT=4e8810b1a8637695171ed346ce68f6984e585ef4
 HLSLIB_COMMIT=cfafe11a93b79ab1af7529d68f08886913a6466e
 PYVERILATOR_COMMIT=c97a5ba41bbc7c419d6f25c74cdf3bdc3393174f

src/finn/custom_op/debugmarker.py

+# Copyright (c) 2020, Xilinx
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+#
+# * Redistributions of source code must retain the above copyright notice, this
+#   list of conditions and the following disclaimer.
+#
+# * Redistributions in binary form must reproduce the above copyright notice,
+#   this list of conditions and the following disclaimer in the documentation
+#   and/or other materials provided with the distribution.
+#
+# * Neither the name of FINN nor the names of its
+#   contributors may be used to endorse or promote products derived from
+#   this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+from finn.custom_op import CustomOp
+from onnx import helper
+
+
+class DebugMarker(CustomOp):
+    def get_nodeattr_types(self):
+        return {"export_debug_name": ("s", True, "")}
+
+    def make_shape_compatible_op(self, model):
+        node = self.onnx_node
+        return helper.make_node("Identity", [node.input[0]], [node.output[0]])
+
+    def infer_node_datatype(self, model):
+        node = self.onnx_node
+        # data type stays the same
+        dtype = model.get_tensor_datatype(node.input[0])
+        model.set_tensor_datatype(node.output[0], dtype)
+        # create quantization annotation for debug marker
+        model.set_tensor_datatype(self.get_nodeattr("export_debug_name"), dtype)
+
+    def execute_node(self, context, graph):
+        node = self.onnx_node
+        inp_name = node.input[0]
+        out_name = node.output[0]
+        inp = context[inp_name]
+        context[out_name] = inp
+        # insert debug marker output as separate tensor
+        context[self.get_nodeattr("export_debug_name")] = inp
+
+    def verify_node(self):
+        info_messages = []
+        # verify that "domain" is set to "finn"
+        domain_value = self.onnx_node.domain
+        if domain_value == "finn":
+            info_messages.append("Attribute domain is set correctly")
+        else:
+            info_messages.append('Attribute domain should be set to "finn"')
+        return info_messages

src/finn/custom_op/registry.py

@@ -57,6 +57,7 @@ from finn.custom_op.fpgadataflow.vector_vector_activate_batch import (
 )
 from finn.custom_op.fpgadataflow.channelwise_op_batch import ChannelwiseOp_Batch
 from finn.custom_op.fpgadataflow.iodma import IODMA
+from finn.custom_op.debugmarker import DebugMarker
 
 # create a mapping of all known CustomOp names and classes
 custom_op = {}
@@ -84,6 +85,7 @@ custom_op["DuplicateStreams_Batch"] = DuplicateStreams_Batch
 custom_op["Vector_Vector_Activate_Batch"] = Vector_Vector_Activate_Batch
 custom_op["ChannelwiseOp_Batch"] = ChannelwiseOp_Batch
 custom_op["IODMA"] = IODMA
+custom_op["DebugMarker"] = DebugMarker
 
 
 def getCustomOp(node):

src/finn/util/pytorch.py

+# Copyright (c) 2020, Xilinx
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+#
+# * Redistributions of source code must retain the above copyright notice, this
+#   list of conditions and the following disclaimer.
+#
+# * Redistributions in binary form must reproduce the above copyright notice,
+#   this list of conditions and the following disclaimer in the documentation
+#   and/or other materials provided with the distribution.
+#
+# * Neither the name of FINN nor the names of its
+#   contributors may be used to endorse or promote products derived from
+#   this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+import torch
+
+from torch.nn import Module, Sequential
+from brevitas.quant_tensor import QuantTensor
+
+
+class Normalize(Module):
+    def __init__(self, mean, std, channels):
+        super(Normalize, self).__init__()
+
+        self.mean = mean
+        self.std = std
+        self.channels = channels
+
+    def forward(self, x):
+        x = x - torch.tensor(self.mean, device=x.device).reshape(1, self.channels, 1, 1)
+        x = x / self.std
+        return x
+
+
+class ToTensor(Module):
+    def __init__(self):
+        super(ToTensor, self).__init__()
+
+    def forward(self, x):
+        x = x / 255
+        return x
+
+
+class NormalizePreProc(Module):
+    def __init__(self, mean, std, channels):
+        super(NormalizePreProc, self).__init__()
+        self.features = Sequential()
+        scaling = ToTensor()
+        self.features.add_module("scaling", scaling)
+        normalize = Normalize(mean, std, channels)
+        self.features.add_module("normalize", normalize)
+
+    def forward(self, x):
+        return self.features(x)
+
+
+class BrevitasDebugHook:
+    def __init__(self):
+        self.outputs = {}
+
+    def __call__(self, module, module_in, module_out):
+        tensor = module_out
+        if isinstance(module_out, QuantTensor):
+            tensor = module_out[0]
+        self.outputs[module.export_debug_name] = tensor.detach().numpy()
+
+    def clear(self):
+        self.outputs = {}

tests/brevitas/test_brevitas_debug.py

+# Copyright (c) 2020, Xilinx
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+#
+# * Redistributions of source code must retain the above copyright notice, this
+#   list of conditions and the following disclaimer.
+#
+# * Redistributions in binary form must reproduce the above copyright notice,
+#   this list of conditions and the following disclaimer in the documentation
+#   and/or other materials provided with the distribution.
+#
+# * Neither the name of FINN nor the names of its
+#   contributors may be used to endorse or promote products derived from
+#   this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+from pkgutil import get_data
+
+import os
+import brevitas.onnx as bo
+import numpy as np
+import onnx
+import onnx.numpy_helper as nph
+import torch
+
+import finn.core.onnx_exec as oxe
+from finn.core.modelwrapper import ModelWrapper
+from finn.transformation.fold_constants import FoldConstants
+from finn.transformation.general import RemoveStaticGraphInputs
+from finn.transformation.infer_shapes import InferShapes
+from finn.util.test import get_test_model_trained
+from finn.util.pytorch import BrevitasDebugHook
+
+
+def test_brevitas_debug():
+    finn_onnx = "test_brevitas_debug.onnx"
+    fc = get_test_model_trained("TFC", 2, 2)
+    dbg_hook = BrevitasDebugHook()
+    bo.enable_debug(fc, dbg_hook)
+    bo.export_finn_onnx(fc, (1, 1, 28, 28), finn_onnx)
+    model = ModelWrapper(finn_onnx)
+    model = model.transform(InferShapes())
+    model = model.transform(FoldConstants())
+    model = model.transform(RemoveStaticGraphInputs())
+    assert len(model.graph.input) == 1
+    assert len(model.graph.output) == 1
+    # load one of the test vectors
+    raw_i = get_data("finn", "data/onnx/mnist-conv/test_data_set_0/input_0.pb")
+    input_tensor = onnx.load_tensor_from_string(raw_i)
+    # run using FINN-based execution
+    input_dict = {"0": nph.to_array(input_tensor)}
+    output_dict = oxe.execute_onnx(model, input_dict, return_full_exec_context=True)
+    produced = output_dict[model.graph.output[0].name]
+    # run using PyTorch/Brevitas
+    input_tensor = torch.from_numpy(nph.to_array(input_tensor)).float()
+    assert input_tensor.shape == (1, 1, 28, 28)
+    # do forward pass in PyTorch/Brevitas
+    expected = fc.forward(input_tensor).detach().numpy()
+    assert np.isclose(produced, expected, atol=1e-3).all()
+    # check all tensors at debug markers
+    names_brevitas = set(dbg_hook.outputs.keys())
+    names_finn = set(output_dict.keys())
+    names_common = names_brevitas.intersection(names_finn)
+    assert len(names_common) == 8
+    for dbg_name in names_common:
+        assert (dbg_hook.outputs[dbg_name] == output_dict[dbg_name]).all()
+    os.remove(finn_onnx)