nshmyrev
commented
5 years ago Thank you for reporting this issue. Please submit a pull request to update the t2t version and remove unused parameter.
Open william-vw opened 5 years ago
nshmyrev
commented
5 years ago Thank you for reporting this issue. Please submit a pull request to update the t2t version and remove unused parameter.
william-vw
commented
5 years ago Ok, done
renan-cunha
commented
5 years ago I run into the same problem and now is appearing:
create_run_config() missing 1 required positional argument: 'model_name'
Is there a stable version?
william-vw
commented
5 years ago I don't think that the pull request has been merged yet - but in the meantime you can simply remove the referred line from g2p_trainer_utils.py.
agadooPineapple
commented
5 years ago I am running in a python 3.5 virtual environment and have the same error as @renan-cunha . After removing line 217 (cpu shard) as suggested and running the test again I have the following output:
test_decode (tests.g2p_unittest.TestG2P) ... ERROR
test_evaluate (tests.g2p_unittest.TestG2P) ... ERROR
test_train (tests.g2p_unittest.TestG2P) ... ERROR
======================================================================
ERROR: test_decode (tests.g2p_unittest.TestG2P)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/tests/g2p_unittest.py", line 34, in test_decode
g2p_model = G2PModel(params, test_path=decode_file_path)
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p.py", line 89, in __init__
self.__prepare_model()
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p.py", line 102, in __prepare_model
self.params)
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p_trainer_utils.py", line 229, in create_run_config
use_tpu=params.use_tpu)
TypeError: create_run_config() missing 1 required positional argument: 'model_name'
======================================================================
ERROR: test_evaluate (tests.g2p_unittest.TestG2P)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/tests/g2p_unittest.py", line 47, in test_evaluate
g2p_model = G2PModel(params, test_path=gt_path)
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p.py", line 89, in __init__
self.__prepare_model()
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p.py", line 102, in __prepare_model
self.params)
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p_trainer_utils.py", line 229, in create_run_config
use_tpu=params.use_tpu)
TypeError: create_run_config() missing 1 required positional argument: 'model_name'
======================================================================
ERROR: test_train (tests.g2p_unittest.TestG2P)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/tests/g2p_unittest.py", line 23, in test_train
test_path=dev_path)
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p.py", line 77, in __init__
self.__prepare_model(train_mode=True)
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p.py", line 102, in __prepare_model
self.params)
File "/home/ash/Downloaded_programs/p3_5s2s/g2p-seq2seq/g2p_seq2seq/g2p_trainer_utils.py", line 229, in create_run_configstuff
use_tpu=params.use_tpu)
TypeError: create_run_config() missing 1 required positional argument: 'model_name'
----------------------------------------------------------------------
Ran 3 tests in 0.005s
FAILED (errors=3)
Test failed: <unittest.runner.TextTestResult run=3 errors=3 failures=0>
error: Test failed: <unittest.runner.TextTestResult run=3 errors=3 failures=0>@william-vw did you need to change anything else to get it to work?
I also have a python 2.7 virtual environment and it gives a different error for each test:
TypeError: create_run_config() takes at least 1 argument (23 given)
Thinking that somewhere where this function is called it is still putting something in for CPU shard and getting too many arguments. Python 3.5 thinks you deleted model_name and that is what is missing, python 2.7 thinks you are missing 1 argument. (Before shard CPU line was deleted there were 24)
agadooPineapple
commented
5 years ago I was using the latest tensorflow and tensor2tensor version 1.12.0 I found downgrading to 1.9.0 for both I don't even have to remove line 217, I still get errors/warnings but 2/3 tests are passing now.
william-vw
commented
5 years ago No, I didn't have to do anything else to get it to work ..
My setup:
mondhs
commented
5 years ago Just to add some details. Hope it helps, I also received errors with 'shard_to_cpu'. After I applied fix manually I got error with 'model_name', when running test. When I check versions:
tensor2tensor 1.12.0
tensorflow 1.12.0
g2p-seq2seq 6.2.2a0 So I downgraded tensor2tensor to 1.6.6 and now tests are passing.
Update 1: Running G2P is working, but training with custom dictionary not. It is showing errors:
File "/usr/local/lib/python3.6/dist-packages/g2p_seq2seq-6.2.2a0-py3.6.egg/g2p_seq2seq/g2p_problem.py", line 449, in create_data_files
save_dic(train_dic, train_path)
File "/usr/local/lib/python3.6/dist-packages/g2p_seq2seq-6.2.2a0-py3.6.egg/g2p_seq2seq/g2p_problem.py", line 520, in save_dic
save_file.write(word + " " + pron + "\n")
TypeError: unsupported operand type(s) for +: 'NoneType' and 'str'Update 2: My bad... My dictionary had the last line empty. I could fix the source, but changed code instead:
diff --git a/g2p_seq2seq/g2p_problem.py b/g2p_seq2seq/g2p_problem.py
index 7943bdb..2bb14bc 100644
--- a/g2p_seq2seq/g2p_problem.py
+++ b/g2p_seq2seq/g2p_problem.py
@@ -516,5 +516,7 @@ def split_graphemes_phonemes(input_line, cleanup=False, p2g_mode=False):
def save_dic(dic, save_path):
with tf.gfile.GFile(save_path, mode="w") as save_file:
for word, pronunciations in dic.items():
+ if(not word):
+ continue
for pron in pronunciations:
save_file.write(word + " " + pron + "\n")
I just tried running the tests using
sudo python setup.py testand got the following error, which is seems related to the fact that theshard_to_cpuparameter was removed a while ago. After removing the referred line ing2p_trainer_utils.pyit works fine.