hanguo97 / multitasksimplification Goto Github PK

how to solve the problem?

Hi! I am trying to train a model using your code, however I am confused about which parameters to use. I am currently getting the error below:

raceback (most recent call last):
File "run.py", line 482, in
tf.app.run()
File "/home/ggarbace/.local/lib/python2.7/site-packages/tensorflow/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "run.py", line 467, in main
FLAGS, hps = add_arguments()
File "run.py", line 204, in add_arguments
raise ValueError("names and train_data_dirs not match")
ValueError: names and train_data_dirs not match

Thanks, and I hope to hear from you.

Hi,
Could you please provide the training time different mtl methods(soft/hard/dynamic etc)?

Hi, I'm unable to load the pre-trained models on Google Colab. I'm getting the following error:
ValueError: Can't load save_path when it is None.
The following code snippet was used:
with tf.Session() as sess:
saver = tf.train.import_meta_graph('My Drive/Colab Notebooks/Multi_Tasking/model-final.meta')
saver.restore(sess, tf.train.latest_checkpoint('My Drive/Colab Notebooks/Multi_Tasking/'))

I don't find the vocab file corresponding to the pre-trained model

What are the pretrained models trained on wikilarge, newsela or wikismall?

Hi,

Would it be possible to get the model output for all the datasets?

Thanks!

I use the command below to sample from the pre-trained model:

    python run.py --mode decode --decode_ckpt_file NewselaFinalMTL/model-final --decode_data_dir decode/test_file.src --decode_output_file output.txt --log_root logs --exp_name test --soft_sharing_coef 0.1 --vocab_path cnn-dailymail/finished_files/vocab --batch_size 32 --beam_size 5 --max_enc_steps 5 --max_dec_steps 5 --steps_per_eval 5 --lr 0.01

However, I get the following error:

Task Newsela is using SummarizationModel
Task SNLI is using SummarizationModel
Task PP is using SummarizationModel
Loading Decoding Data from MultitaskSimplification2/decode/test_file.src
max_size of vocab was specified as 50000; we now have 50000 words. Stopping reading.
Finished constructing vocabulary of 50000 total words. Last word added: 18:19
Exception in thread Thread-1:
Traceback (most recent call last):
File "/home/usr/anaconda2/lib/python2.7/threading.py", line 801, in __bootstrap_inner
self.run()
File "/home/usr/anaconda2/lib/python2.7/threading.py", line 754, in run
self.__target(*self.__args, **self.__kwargs)
File "MultitaskSimplification/pointer_model/batcher.py", line 341, in fill_example_queue
(article, abstract) = input_gen.next()
File "MultitaskSimplification/pointer_model/batcher.py", line 423, in text_generator
e = example_generator.next() # e is a tf.Example
File "MultitaskSimplification/pointer_model/data.py", line 139, in example_generator
example_str = struct.unpack('%ds' % str_len, reader.read(str_len))[0]
MemoryError

Creating SummarizationModel Newsela Model
INFO:tensorflow:Building graph...
TODO: Using tf.where to replace tf.cond in next_cell_input
change sampled_tokens not include ?
INFO:tensorflow:Time to build graph: 2 seconds
2020-05-05 14:35:23.921164: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
INFO:tensorflow:Restoring parameters from MultitaskSimplification/NewselaFinalMTL/model-final
Traceback (most recent call last):
File "run.py", line 482, in
tf.app.run()
File "/home/usr/anaconda2/lib/python2.7/site-packages/usr/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "run.py", line 475, in main
setup_and_run_decoding(FLAGS, hps)
File "run.py", line 452, in setup_and_run_decoding
decoder.decode(ckpt_file=FLAGS.decode_ckpt_file)
File "MultitaskSimplification/pointer_model/decode.py", line 63, in decode
misc_utils.load_ckpt(self._saver, self._sess, self._ckpt_dir, ckpt_file)
File "MultitaskSimplification/utils/misc_utils.py", line 64, in load_ckpt
saver.restore(sess, ckpt_file)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 1666, in restore
{self.saver_def.filename_tensor_name: save_path})
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 889, in run
run_metadata_ptr)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1120, in _run
feed_dict_tensor, options, run_metadata)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1317, in _do_run
options, run_metadata)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1336, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Assign requires shapes of both tensors to match. lhs shape= [50000] rhs shape= [29131]
[[Node: save_1/Assign_11 = Assign[T=DT_FLOAT, _class=["loc:@Newsela_Attention/policy/rnn/logits_kernel/bias"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"](Newsela_Attention/policy/rnn/logits_kernel/bias, save_1/RestoreV2_11)]]

Caused by op u'save_1/Assign_11', defined at:
File "run.py", line 482, in
tf.app.run()
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "run.py", line 475, in main
setup_and_run_decoding(FLAGS, hps)
File "run.py", line 449, in setup_and_run_decoding
decoder.build_graph(decode_sess)
File "/srv/disk01/usr/TS/models/MultitaskSimplification/pointer_model/decode.py", line 57, in build_graph
self._saver = tf.train.Saver()
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 1218, in init
self.build()
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 1227, in build
self._build(self._filename, build_save=True, build_restore=True)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 1263, in _build
build_save=build_save, build_restore=build_restore)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 751, in _build_internal
restore_sequentially, reshape)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 439, in _AddRestoreOps
assign_ops.append(saveable.restore(tensors, shapes))
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 160, in restore
self.op.get_shape().is_fully_defined())
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/ops/state_ops.py", line 276, in assign
validate_shape=validate_shape)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/ops/gen_state_ops.py", line 57, in assign
use_locking=use_locking, name=name)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper
op_def=op_def)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 2956, in create_op
op_def=op_def)
File "/home/usr/anaconda2/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 1470, in init
self._traceback = self._graph._extract_stack() # pylint: disable=protected-access

InvalidArgumentError (see above for traceback): Assign requires shapes of both tensors to match. lhs shape= [50000] rhs shape= [29131]
[[Node: save_1/Assign_11 = Assign[T=DT_FLOAT, _class=["loc:@Newsela_Attention/policy/rnn/logits_kernel/bias"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"](Newsela_Attention/policy/rnn/logits_kernel/bias, save_1/RestoreV2_11)]]