joezhouwenxuan / self-regulation-employing-a-generative-adversarial-network-to-improve-event-detection Goto Github PK

In section 4.5 of the paper Ldiff is combined to the softmax loss using lambda to optimize Θd. As I understand it, it should be applied to Θg, not Θd since Ldiff is only a function of og and oghat
In the code there are two lambdas:
self.train_op_g = optimizer_g.minimize(self.g_loss + 0.1 * self.diff_loss, var_list=vars_g) self.total_loss = self.loss + self.l2_loss + self.diff_loss * 0.00001
and the diff loss is applied to g, not d.
Is the code the proper version ?

File "train.py", line 77, in
main()
File "train.py", line 58, in main
model = Model(config)
File "/home/zhou/文档/Self-regulation-Employing-a-Generative-Adversarial-Network-to-Improve-Event-Detection-master/model.py", line 42, in init
self.build()
File "/home/zhou/文档/Self-regulation-Employing-a-Generative-Adversarial-Network-to-Improve-Event-Detection-master/model.py", line 78, in build
name='feature_embedding'
File "/home/zhou/.local/lib/python3.6/site-packages/tensorflow/python/ops/embedding_ops.py", line 122, in embedding_lookup
return maybe_normalize(_do_gather(params[0], ids, name=name))
File "/home/zhou/.local/lib/python3.6/site-packages/tensorflow/python/ops/embedding_ops.py", line 42, in _do_gather
return array_ops.gather(params, ids, name=name)
File "/home/zhou/.local/lib/python3.6/site-packages/tensorflow/python/ops/gen_array_ops.py", line 1179, in gather
validate_indices=validate_indices, name=name)
File "/home/zhou/.local/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py", line 767, in apply_op
op_def=op_def)
File "/home/zhou/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 2506, in create_op
original_op=self._default_original_op, op_def=op_def)
File "/home/zhou/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 1269, in init
self._traceback = _extract_stack()

InvalidArgumentError (see above for traceback): indices[0,0] = 531 is not in [0, 26)
[[Node: feature_embedding = Gather[Tindices=DT_INT32, Tparams=DT_FLOAT, _class=["loc:@feature_W"], validate_indices=true, _device="/job:localhost/replica:0/task:0/cpu:0"](feature_W/read, _arg_input_0_1)]]

I have the ACE 2005 corpus and in this code, it seems that the train.tks and train.tgs is just an example? I don't really understand how to generate these two files.
Besides, from my understanding, I can use the wordlist and labellist in the data folder directly but have to generate another wordvec file, am I right?

How to make sure the generator of GAN does not generate features randomly, that it can maximum the loss L(prediction, ground truth)?