FIX: lm pre-training #58
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
finetune/base.py
Outdated
| @@ -489,6 +493,8 @@ def _build_model(self, n_updates_total, target_dim, train=True): | |||
| elif not self.is_trained: | |||
| self._load_base_model() | |||
|
|
|||
| guarantee_initialized_variables(self.sess) | |||
There was a problem hiding this comment.
Can you explain the thought process behind this? Can't we explicitly initialize the variables under model/clf instead? Or add some logs for the variables that this call initializes. I can foresee a future where we accidentally mess up the base model and this could hide problems?
There was a problem hiding this comment.
Specify model/clf as scope
finetune/base.py
Outdated
| self.do_dropout = tf.placeholder(tf.float32) # 1 for do dropout and 0 to not do dropout | ||
| if self.target_type == SEQUENCE_LABELING: | ||
| self.Y = tf.placeholder(tf.int32, [None, self.config.max_length]) # classification targets | ||
| else: | ||
| self.Y = tf.placeholder(tf.float32, [None, self.target_dim]) # classification targets | ||
| self.Y = tf.stop_gradient(tf.placeholder(tf.float32, [None, self.target_dim or 1])) # classification targets |
There was a problem hiding this comment.
Gradients already do not flow into placeholders The tf.stop_gradient should be on the loss function to protect against changes to self.Y
finetune/utils.py
Outdated
| @@ -95,6 +95,14 @@ def np_init(w): | |||
| return partial(_np_init, w=w) | |||
|
|
|||
|
|
|||
| def guarantee_initialized_variables(sess): | |||
There was a problem hiding this comment.
I really like this.
madisonmay
commented
Jul 23, 2018
finetune/utils.py
Outdated
| @@ -95,6 +95,14 @@ def np_init(w): | |||
| return partial(_np_init, w=w) | |||
|
|
|||
|
|
|||
| def guarantee_initialized_variables(sess): | |||
| global_vars = tf.global_variables() | |||
There was a problem hiding this comment.
Credit stackoverflow
madisonmay
force-pushed
the
madison/lm-pretraining
branch
from
0a7f620
to
ae3823d
Compare
madisonmay
force-pushed
the
madison/lm-pretraining
branch
from
0ab0fa0
to
0988082
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Resolves #56