Skip to content
Snippets Groups Projects
Commit 61f39c45 authored by Yaman Umuroglu's avatar Yaman Umuroglu
Browse files

[Test] reorganize test folders

parent 119bbca2
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Showing
with 10 additions and 136 deletions
tests/test_basic_onnx_exec.py → tests/core/test_basic_onnx_exec.py
+ 0
− 0
View file @ 61f39c45
File moved
tests/test_mixed_onnx_exec.py → tests/core/test_mixed_onnx_exec.py
+ 0
− 0
View file @ 61f39c45
File moved
tests/test_finn_hls_onnx_graph.py deleted 100644 → 0
+ 0
− 128
View file @ 119bbca2
import onnx
from onnx import TensorProto, helper
from finn.core.datatype import DataType
from finn.core.modelwrapper import ModelWrapper
def test_manually_construct_onnx_graph():
inp = helper.make_tensor_value_info("inp", TensorProto.FLOAT, [1, 13, 64])
outp = helper.make_tensor_value_info("outp", TensorProto.FLOAT, [1, 1, 64])
memInStrm_node = helper.make_node(
"FIFO", ["inp"], ["memInStrm"], "memInStrm", domain='finn', backend='fpgadataflow', depth=1024
)
FCLayer0_node = helper.make_node(
"StreamingFCLayer_Batch",
["memInStrm", "weights0", "thresh0"],
["out1"],
domain='finn',
backend='fpgadataflow',
resType="ap_resource_lut()",
MW=832,
MH=1024,
SIMD=64,
PE=32,
resDataType="Recast<XnorMul>",
)
inter0_node = helper.make_node("FIFO", ["out1"], ["inter0"], "inter0", domain='finn', backend='fpgadataflow', depth=16)
FCLayer1_node = helper.make_node(
"StreamingFCLayer_Batch",
["inter0", "weights1", "thresh1"],
["out2"],
domain='finn',
backend='fpgadataflow',
resType="ap_resource_lut()",
MW=1024,
MH=1024,
SIMD=32,
PE=64,
resDataType="Recast<XnorMul>",
)
inter1_node = helper.make_node("FIFO", ["out2"], ["inter1"], "inter1", domain='finn', backend='fpgadataflow', depth=16)
FCLayer2_node = helper.make_node(
"StreamingFCLayer_Batch",
["inter1", "weights2", "thresh2"],
["out3"],
domain='finn',
backend='fpgadataflow',
resType="ap_resource_lut()",
MW=1024,
MH=1024,
SIMD=64,
PE=32,
resDataType="Recast<XnorMul>",
)
inter2_node = helper.make_node("FIFO", ["out3"], ["inter2"], "inter2", domain='finn', backend='fpgadataflow', depth=8)
FCLayer3_node = helper.make_node(
"StreamingFCLayer_Batch",
["inter2", "weights3", "thresh3"],
["out4"],
domain='finn',
backend='fpgadataflow',
resType="ap_resource_lut()",
MW=1024,
MH=64,
SIMD=8,
PE=16,
resDataType="Recast<XnorMul>",
)
memOutStrm_node = helper.make_node(
"FIFO", ["out4"], ["outp"], "memOutStrm", domain='finn', backend='fpgadataflow', depth=1024
)
graph = helper.make_graph(
nodes=[
memInStrm_node,
FCLayer0_node,
inter0_node,
FCLayer1_node,
inter1_node,
FCLayer2_node,
inter2_node,
FCLayer3_node,
memOutStrm_node,
],
name="finn_hls_onnx_graph",
inputs=[inp],
outputs=[outp],
value_info=[
helper.make_tensor_value_info("memInStrm", TensorProto.FLOAT, [1, 13, 64]),
helper.make_tensor_value_info("weights0", TensorProto.FLOAT, [64, 32, 416]),
helper.make_tensor_value_info(
"thresh0", TensorProto.FLOAT, [32, 32, 1, 16, 1]
),
helper.make_tensor_value_info("out1", TensorProto.FLOAT, [1, 32, 32]),
helper.make_tensor_value_info("inter0", TensorProto.FLOAT, [1, 32, 32]),
helper.make_tensor_value_info("weights1", TensorProto.FLOAT, [32, 64, 512]),
helper.make_tensor_value_info(
"thresh1", TensorProto.FLOAT, [16, 64, 1, 16, 1]
),
helper.make_tensor_value_info("out2", TensorProto.FLOAT, [1, 16, 64]),
helper.make_tensor_value_info("inter1", TensorProto.FLOAT, [1, 16, 64]),
helper.make_tensor_value_info("weights2", TensorProto.FLOAT, [64, 32, 512]),
helper.make_tensor_value_info(
"thresh2", TensorProto.FLOAT, [32, 32, 1, 16, 1]
),
helper.make_tensor_value_info("out3", TensorProto.FLOAT, [1, 32, 32]),
helper.make_tensor_value_info("inter2", TensorProto.FLOAT, [1, 32, 32]),
helper.make_tensor_value_info("weights3", TensorProto.FLOAT, [8, 16, 512]),
helper.make_tensor_value_info(
"thresh3", TensorProto.FLOAT, [4, 16, 1, 16, 1]
),
helper.make_tensor_value_info("out4", TensorProto.FLOAT, [1, 1, 64]),
],
)
model = helper.make_model(graph, producer_name="finn-hls-onnx-model")
model = ModelWrapper(model)
# set the tensor datatypes (in this case: all to bipolar)
for tensor in graph.input:
model.set_tensor_datatype(tensor.name, DataType["BIPOLAR"])
for tensor in graph.output:
model.set_tensor_datatype(tensor.name, DataType["BIPOLAR"])
for tensor in graph.value_info:
model.set_tensor_datatype(tensor.name, DataType["BIPOLAR"])
#onnx.save(model.model, "finn-hls-onnx-model.onnx")
tests/test_padding.py → tests/util/test_padding.py
+ 0
− 0
View file @ 61f39c45
File moved
tests/fpgadataflow/test_rtlsim2npy.py → tests/util/test_rtlsim2npy.py
+ 10
− 8
View file @ 61f39c45
......@@ -47,7 +47,9 @@ def test_unpack_innermost_dim_from_hex_string():
dtype = DataType.INT2
shape = (1, 2, 2, 2)
eC = [[[-1, -1], [-1, -1]], [[-1, 1], [1, -1]]]
C_unpacked = unpack_innermost_dim_from_hex_string(C, dtype, shape, reverse_inner=True)
C_unpacked = unpack_innermost_dim_from_hex_string(
C, dtype, shape, reverse_inner=True
)
assert (C_unpacked == eC).all()
# INT4
......@@ -58,20 +60,20 @@ def test_unpack_innermost_dim_from_hex_string():
D_unpacked = unpack_innermost_dim_from_hex_string(D, dtype, shape)
assert (D_unpacked == eD).all()
D_unpacked = unpack_innermost_dim_from_hex_string(D, dtype, shape, reverse_inner=True)
D_unpacked = unpack_innermost_dim_from_hex_string(
D, dtype, shape, reverse_inner=True
)
assert (D_unpacked == eD).all()
# INT32
E = np.asarray(["0xffffffff", "0xfffffffe", "0x02", "0xffffffef"])
dtype = DataType.INT32
shape = (1, 4, 1)
eE = [[[-1],[-2],[2],[-17]]]
eE = [[[-1], [-2], [2], [-17]]]
E_unpacked = unpack_innermost_dim_from_hex_string(E, dtype, shape)
assert (E_unpacked == eE).all()
E_unpacked = unpack_innermost_dim_from_hex_string(E, dtype, shape, reverse_inner=True)
E_unpacked = unpack_innermost_dim_from_hex_string(
E, dtype, shape, reverse_inner=True
)
assert (E_unpacked == eE).all()
tests/test_shape_utils.py → tests/util/test_shape_utils.py
+ 0
− 0
View file @ 61f39c45
File moved
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment