Could this be adapted to use multiple score features instead of a single score?

Hi,

Got below error while reading the training data:
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x92 in position 80: invalid start byte

Tried all the ways, but could not read the file,
Finally converted the file from tsv to txt , and could read it( but I feel this is a workaround rather than a solution)

Got similar error while reading the glove file.
UnicodeDecodeError: 'charmap' codec can't decode byte 0x9d in position 2438: character maps to

Please provide help on this.

Regards,
Pooja

There is no rate parameter in tf.nn.drop for tensorflow 1.10

After running python cv_train.py --essay_set_id 1 in terminal, I get this error:

Traceback (most recent call last):
  File "cv_train.py", line 35, in <module>
    FLAGS._parse_flags()
  File "/home/aruhi/tensorflow/venv/lib/python3.6/site-packages/tensorflow/python/platform/flags.py", line 85, in __getattr__
    return wrapped.__getattr__(name)
  File "/home/aruhi/tensorflow/venv/lib/python3.6/site-packages/absl/flags/_flagvalues.py", line 470, in __getattr__
    raise AttributeError(name)
AttributeError: _parse_flags

Hello, I am confused about the variable logits_bias. It seems that you make logits_bias not trainable in your code. (At line 170 in memn2n_kv.py)

Other possible mistakes I found:
1.

Thanks!

Followed the instructions. Using:
python2.7
tensorflow 1.3
Ubuntu 16.04
sklean 19.1

I can show the versions of the rest of the packages if that would help.

Got this error:

Traceback (most recent call last):
  File "cv_train.py", line 245, in <module>
    sess.run(tf.global_variables_initializer(), feed_dict={model.w_placeholder: word2vec})
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 766, in run
    run_metadata_ptr)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 937, in _run
    np_val = np.asarray(subfeed_val, dtype=subfeed_dtype)
  File "/usr/local/lib/python2.7/dist-packages/numpy/core/numeric.py", line 531, in asarray
    return array(a, dtype, copy=False, order=order)
ValueError: setting an array element with a sequence.

any suggestions would be helpful.

why are the parameter values displayed as below
ALLOW_SOFT_PLACEMENT=<absl.flags._flag.BooleanFlag object at 0x7f888d8f1610>

Got the following error while trying to run cv_train.py
E = data_utils.vectorize_data(essay_list, word_idx, max_sent_size)
^
SyntaxError: invalid syntax

I've come back to this repo after some time and started testing again.

Some epoch folds are resulting in kappa scores of 0.

Is this know or expected?

The paper 'A Memory-Augmented Neural Model for Automated Grading' demonstrates some very impressive accuracy scores.

https://siyuanzhao.github.io/pdf/L_S_2017.pdf

I'm trying to reproduce those results.

What parameter and flag values do I need to set to reproduce the research findings?

Hi Authors,

Is there a procedure where I could save the weights of the model. Tried playing around but was not able to.

Please let me know at the earliest.