def __init__(self, func):
self.func = func
src = clip_head(inspect.getsource(func))
dprint(src)
self.ast = gast.ast_to_gast(ast.parse(src)).body[0]
assert (isinstance(self.ast, gast.gast.FunctionDef))
def __init__(self, num_hidden):
super(LinkInFor, self).__init__()
with self.init_scope():
self.l = L.Linear(num_hidden, num_hidden)
def forward(self, x, h, indices):
for i in indices:
h = h + self.l(x[:, i])
return h
import ch2o
if __name__ == '__main__':
import numpy as np
np.random.seed(314)
batch_size = 3
num_hidden = 5
sequence_length = 4
model = LinkInFor(num_hidden)
x = np.random.rand(
batch_size, sequence_length, num_hidden).astype(np.float32)
h = np.random.rand(batch_size, num_hidden).astype(np.float32)
args = [x, h, np.arange(sequence_length)]
dprint(model(*args))
ch2o.generate_testcase(model, args)
def compile_model(model, inputs):
# return helper.make_graph([],'dummy',[],[])
init_id2name(model)
# code.InteractiveConsole({'mo': model}).interact()
env = Env(sys.modules[model.__module__])
molk = User_Defined_Link(model, env)
input_tensors = []
for i in inputs:
# TODO(hamaji): Set valid type info.
if isinstance(i, (list, tuple)):
x = new_sequence()
elif i is None:
x = new_tensor()
else:
if isinstance(i, int):
i = np.array(i)
else:
# TODO(durswd): This code requires chainer6.x
i = chainer.cuda.to_cpu(i)
x = new_tensor(dims=i.shape, dtype=i.dtype)
input_tensors.append(x)
input_values = [Value(i) for i in input_tensors]
v = molk.call(input_values, [], env)
dprint('output_tensors', v)
if isinstance(v.value, tuple):
output_tensors = list(v.value) # ばらしてみる
else:
output_tensors = [v] # とりあえず1tensor
# print('env.init_tensors ',env.init_tensors)
input_tensors += list(env.init_tensors.values())
for f in env.restore_funcs:
f()
# for no in env.nodes:
# print(no.op_type)
# print(env.nodes)
# print(input_tensors)
# print(output_tensors)
# for ch in model.namedparams():
# print(ch)
outputs_vi = [o.to_value_info(env) for o in output_tensors]
graph = make_graph(env.nodes, 'name_is_unknown_now', input_tensors,
outputs_vi)
# inputのうち、重みであるものにはinitializerをつける
# batch_sizeやinput_sizeなどの可変なものはできる限りのそのままで
# Chainer compiler 独自のノードを使うとcheckできなくなる...
# checker.check_graph(graph)
mo = helper.make_model(graph)
# print(mo)
return mo