search

Sshanu / Relation-Classification-using-Bidirectional-LSTM-Tree

TensorFlow Implementation of the paper "End-to-End Relation Extraction using LSTMs on Sequences and Tree Structures" and "Classifying Relations via Long Short Term Memory Networks along Shortest Dependency Paths" for classifying relations
MIT License
185 stars 41 forks source link

Invalid Argument Error #9

Open arunzz opened 6 years ago

arunzz commented 6 years ago

Hi Shanu

I tried Model3v1 for 10 samples. Accordingly I generated train_paths and when i run, I get the below error. PLease let know. I have attached the source file too - model3v1.txt

File "", line 1, in [elided 1 identical lines from previous traceback] File "/usr/lib/python2.7/idlelib/run.py", line 324, in runcode exec code in self.locals File "/home/arunkumar/Downloads/Relation-Classification-master/model3v1.py", line 422, in s_seq.get_shape(), tf.TensorShape([None, 100])]) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/control_flow_ops.py", line 3096, in while_loop result = loop_context.BuildLoop(cond, body, loop_vars, shape_invariants) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/control_flow_ops.py", line 2874, in BuildLoop pred, body, original_loop_vars, loop_vars, shape_invariants) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/control_flow_ops.py", line 2814, in _BuildLoop body_result = body(*packed_vars_for_body) File "/home/arunkumar/Downloads/Relation-Classification-master/model3v1.py", line 415, in matmul_hl temp = tf.tanh(tf.matmul(input_seq[j], W) + b) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/array_ops.py", line 588, in _slice_helper name=name) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/array_ops.py", line 754, in strided_slice shrink_axis_mask=shrink_axis_mask) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/gen_array_ops.py", line 5397, in strided_slice name=name) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper op_def=op_def) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 3271, in create_op op_def=op_def)

InvalidArgumentError (see above for traceback): slice index 1 of dimension 0 out of bounds. [[Node: gradients_1/hidden_layer_seq/while_3/strided_slice_grad/StridedSliceGrad = StridedSliceGrad[Index=DT_INT32, T=DT_FLOAT, begin_mask=0, ellipsis_mask=0, end_mask=0, new_axis_mask=0, shrink_axis_mask=1, _device="/job:localhost/replica:0/task:0/device:CPU:0"](gradients_1/hidden_layer_seq/while_3/strided_slice_grad/StridedSliceGrad/StackPopV2, gradients_1/hidden_layer_seq/while_3/strided_slice_grad/StridedSliceGrad/StackPopV2_1, gradients_1/hidden_layer_seq/while_3/strided_slice_grad/StridedSliceGrad/StackPopV2_2, gradients_1/hidden_layer_seq/while_3/strided_slice_grad/StridedSliceGrad/Const_3, gradients_1/hidden_layer_seq/while_3/MatMul_grad/tuple/control_dependency)]]

arunzz commented 6 years ago

Exact error and the code line InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'input/sp_childs' with dtype int32 and shape [5,2,20,20]

For complete model training ---> 29 _, _loss, step, _summary = sess.run([train_op, total_loss, global_step_dep, summary], feed_dict)

Sshanu commented 6 years ago

first check whether [[childs_path1[k], childs_path2[k]] for k in range(s, end)] has the same dimension as [5, 2, 20, 20] and also the data type

arunzz commented 6 years ago

Thank you Shanu for the response. I tried for a small sample as it was slow. But now i tried with all 8000 and am running them in python2 All pickles are written with python2

Now I get this error. InvalidArgumentError (see above for traceback): Inputs to operation gradients_1/AddN1101 of type AddN must have the same size and shape. Input 0: [2,20,1,100] != input 1: [2,1,1,100 This is in model3v2, this line of code For complete model training , _loss, step, _summary = sess.run([train_op, total_loss, global_step_dep, summary], feed_dict)

Kindly help

Sshanu commented 6 years ago

please try using python 3.

Just use printing the shapes of this tensors, then debug the code

NeverInAsh commented 6 years ago

Hey shanu, why should there be a change in dimensions of the tensors, if i am using the same data as yours @arunzz were you able to solve the problem?

Sshanu commented 6 years ago

Can you please post the error or for which tensors dimension problem is occurring?

arunzz commented 6 years ago

Yes Use python 3 It worked. No changes to any dimensions. But the label encoding also has to be passed into relation encoder. That part is missing i think

On Jun 15, 2018, at 11:33 AM, Shanu Kumar notifications@github.com wrote:

Can you please post the error or for which tensors dimension problem is occurring?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

NeverInAsh commented 6 years ago

@arunzz can you share some more insights, actually i am currently learning about it meanwhile i am working on this. It will be a great help. or maybe you can share your version.

arunzz commented 6 years ago

Hi

Shanu is our guru :) He wrote the entire thing.

Regards Arun

On Fri, Jun 15, 2018 at 2:51 PM, NeverInAsh notifications@github.com wrote:

@arunzz https://github.com/arunzz can you share some more insights, actually i am currently learning about it meanwhile i am working on this. It will be a great help. or maybe you can share your version.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Sshanu/Relation-Classification/issues/9#issuecomment-397564692, or mute the thread https://github.com/notifications/unsubscribe-auth/AUJLxe8vgjyeDesR7_i7C6nubFghBQFSks5t83ySgaJpZM4ThR3G .

NeverInAsh commented 6 years ago

It worked, thanks shanu :)

ProxyCausal commented 6 years ago

@arunzz @NeverInAsh I'm wondering what you guys did to fix the error? I am running python 3

arunzz commented 6 years ago

Just python3 Initialize global variables after loading glove embeddings.

On Jul 2, 2018, at 7:59 PM, enmatics notifications@github.com wrote:

@arunzz @NeverInAsh I'm wondering what you guys did to fix the error? I am running python 3

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

zy158 commented 5 years ago

@arunzz I have meet the same problem. I am running python 3, but the problem still exist. Can you attached the worked model3v2 source file?

arunzz commented 5 years ago

model3v2.py.zip

zy158 commented 5 years ago

@arunzz which version of tensorlfow and python you use?

zy158 commented 5 years ago

@arunzz when I don't init glove embedding in your source file, it will exist the error: InvalidArgumentError (see above for traceback): slice index 1 of dimension 0 out of bounds. [[Node: gradients_1/hidden_layer_seq/while_3/strided_slice_grad/StridedSliceGrad = StridedSliceGrad[Index=DT_INT32, T=DT_FLOAT, begin_mask=0, ellipsis_mask=0, end_mask=0, new_axis_mask=0, shrink_axis_mask=1, _device="/job:localhost/replica:0/task:0/device:CPU:0"]

when I init glove embedding, it will exist the error: InvalidArgumentError (see above for traceback): Inputs to operation gradients_1/AddN_1101 of type AddN must have the same size and shape. Input 0: [2,20,1,100] != input 1: [2,1,1,100]

zy158 commented 5 years ago

@arunzz When I use tensorflow 1.12.0, it will lead "ArithmeticOptimize" error. Then I replaced the version of tensorflow 1.8.0. and worked. But it will occur "concat" error and "hidden_layer" error. Then I replaced the version of tensorflow 1.10.0. Your source file finally worked.

radhikagupta6 commented 4 years ago

@arunzz in model3v2 model = tf.train.latest_checkpoint(model_dir) saver.restore(sess, model) when I try to restore the error says 👎 ValueError: Can't load save_path when it is None. Please help