From 80e3d258f19961a0c228189e2163d8fcc2ba988a Mon Sep 17 00:00:00 2001
From: auphelia <jakobapk@web.de>
Date: Wed, 27 May 2020 16:16:18 +0100
Subject: [PATCH] [Transformation] Change creation of sparse matrix for
lowering dw conv to matmul
---
.../transformation/lower_convs_to_matmul.py | 49 ++++++-------------
1 file changed, 16 insertions(+), 33 deletions(-)
diff --git a/src/finn/transformation/lower_convs_to_matmul.py b/src/finn/transformation/lower_convs_to_matmul.py
index 4770b90da..b2831a319 100644
--- a/src/finn/transformation/lower_convs_to_matmul.py
+++ b/src/finn/transformation/lower_convs_to_matmul.py
@@ -27,7 +27,6 @@
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
import numpy as np
-from scipy import sparse
from onnx import TensorProto
from onnx import helper
@@ -64,39 +63,23 @@ class LowerConvsToMatMul(Transformation):
ifm_dim = model.get_tensor_shape(n.input[0])[-1] # assume NCHW
ofm_dim = model.get_tensor_shape(n.output[0])[-1] # assume NCHW
- if group == 1:
- # reuse conv weights for new matmul weights
- # conv weights are [OFM][IFM][k][k]
- # first convert to [OFM][k][k][IFM] (to remain compatible with
- # finn-hlslib and how it does im2col/sliding window)
- W_matmul = W_conv.transpose(0, 2, 3, 1)
- # reshape into [OFM][k*k*IFM] matrix
- W_matmul = W_matmul.reshape(ofm_ch, ifm_ch * k * k)
- # transpose to get ONNX-compatible [k*k*IFM][OFM] matrix
- W_matmul = W_matmul.T
- model.set_initializer(weight_name, W_matmul)
- elif group == ifm_ch and ofm_ch == ifm_ch:
- W_matmul = W_conv.transpose(0, 2, 3, 1)
- # reshape into [OFM][k*k*IFM] matrix
- W_matmul = W_matmul.reshape(ofm_ch, 1 * k * k)
- # transpose to get ONNX-compatible [1*k*k][OFM] matrix
- W_matmul = W_matmul.T
- # create sparse matrix with shape [k*k*IFM][OFM]
- inds = []
- for i in range(k * k):
- for ch in range(ifm_ch):
- inds.append(ch)
- inds = np.array(inds)
- index = np.arange(len(W_matmul.flatten()))
- W_sparse = sparse.csc_matrix(
- (W_matmul.flatten(), (inds, index)),
- shape=(ofm_ch, ifm_ch * k * k),
- )
- W_sparse = W_sparse.toarray().T
- model.set_initializer(weight_name, W_sparse)
+ # if depthwise conv create sparse matrix
+ if group == ifm_ch and ofm_ch == ifm_ch:
+ W_sparse = np.zeros((ofm_ch, ifm_ch, k, k))
+ for ch in range(ifm_ch):
+ W_sparse[ch][ch] = W_conv[ch][0]
+ W_conv = W_sparse.astype(np.float32)
- else:
- raise Exception("Settings for convolution layer not yet supported.")
+ # reuse conv weights for new matmul weights
+ # conv weights are [OFM][IFM][k][k]
+ # first convert to [OFM][k][k][IFM] (to remain compatible with
+ # finn-hlslib and how it does im2col/sliding window)
+ W_matmul = W_conv.transpose(0, 2, 3, 1)
+ # reshape into [OFM][k*k*IFM] matrix
+ W_matmul = W_matmul.reshape(ofm_ch, ifm_ch * k * k)
+ # transpose to get ONNX-compatible [k*k*IFM][OFM] matrix
+ W_matmul = W_matmul.T
+ model.set_initializer(weight_name, W_matmul)
# create new intermediate values
inp_trans_out = helper.make_tensor_value_info(
--
GitLab