From 3e24b974e8f8a34763bbccd7ae867fd0aa8ac532 Mon Sep 17 00:00:00 2001
From: Yaman Umuroglu <yamanu@xilinx.com>
Date: Wed, 22 Sep 2021 10:29:53 +0200
Subject: [PATCH] [Driver] use instance_name, fix ext.w. handling
---
.../fpgadataflow/make_pynq_driver.py | 28 +++++++++++++------
1 file changed, 20 insertions(+), 8 deletions(-)
diff --git a/src/finn/transformation/fpgadataflow/make_pynq_driver.py b/src/finn/transformation/fpgadataflow/make_pynq_driver.py
index 6fa8eb871..bfa2fdbf9 100644
--- a/src/finn/transformation/fpgadataflow/make_pynq_driver.py
+++ b/src/finn/transformation/fpgadataflow/make_pynq_driver.py
@@ -147,7 +147,7 @@ class MakePYNQDriver(Transformation):
ishape_normal.append(i_tensor_shape_normal)
ishape_folded.append(i_tensor_shape_folded)
ishape_packed.append(i_tensor_shape_packed)
- idma_names.append("idma%d" % idma_ind)
+ idma_names.append(getCustomOp(i_consumer).get_nodeattr("instance_name"))
odt = []
odma_names = []
@@ -194,7 +194,7 @@ class MakePYNQDriver(Transformation):
oshape_normal.append(o_tensor_shape_normal)
oshape_folded.append(o_tensor_shape_folded)
oshape_packed.append(o_tensor_shape_packed)
- odma_names.append("odma%d" % odma_ind)
+ odma_names.append(getCustomOp(o_producer).get_nodeattr("instance_name"))
# generate external weights npy files
weights_dir = pynq_driver_dir + "/runtime_weights"
@@ -208,19 +208,31 @@ class MakePYNQDriver(Transformation):
node.op_type == "StreamingDataflowPartition"
), "CreateDataflowPartition needs to be applied before driver generation"
- producer = model.find_producer(node.input[0])
- init_tensor = model.get_initializer(node.input[0])
+ if len(node.input) > 0:
+ producer = model.find_producer(node.input[0])
+ init_tensor = model.get_initializer(node.input[0])
+ else:
+ producer = None
+ init_tensor = None
if producer is None: # input dma?
- idma_name = "idma" + str(idma_idx)
- if init_tensor is not None: # input weights dma?
+ sdp_inst = getCustomOp(node)
+ idma_name = sdp_inst.get_nodeattr("instance_name")
+ df_model = ModelWrapper(sdp_inst.get_nodeattr("model"))
+ assert df_model.graph.node[0].op_type == "IODMA"
+ iodma_node = getCustomOp(df_model.graph.node[0])
+ if iodma_node.get_nodeattr("burstMode") == "wrap": # input weights dma?
+ init_tensor = df_model.get_initializer(
+ iodma_node.onnx_node.input[0]
+ )
ext_weight_dma_cnt += 1
- w_dtype = model.get_tensor_datatype(node.input[0])
+ w_dtype = df_model.get_tensor_datatype(
+ iodma_node.onnx_node.input[0]
+ )
init_external_tensor = to_external_tensor(init_tensor, w_dtype)
np.save(
weights_dir + "/" + idma_name + ".npy", init_external_tensor
)
-
idma_idx += 1
# fill in the driver template
--
GitLab