Welcome to mirror list, hosted at ThFree Co, Russian Federation.

github.com/torch/torch.github.io.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authornicholas-leonard <nick@nikopia.org>2016-07-25 22:38:48 +0300
committernicholas-leonard <nick@nikopia.org>2016-07-25 22:38:48 +0300
commit3aac64502c9083d7db4c6c0b6c1397eef5d9ba76 (patch)
tree2710d53720798590228e2ae8a7eb7949738be5f7
parent7992ccb23ef065bb994ca36ad0eb973c15da273f (diff)
fixed image
-rw-r--r--blog/_posts/2016-05-11-nce.md8
1 files changed, 4 insertions, 4 deletions
diff --git a/blog/_posts/2016-05-11-nce.md b/blog/_posts/2016-05-11-nce.md
index 8a7a3ea..70a6bcb 100644
--- a/blog/_posts/2016-05-11-nce.md
+++ b/blog/_posts/2016-05-11-nce.md
@@ -4,7 +4,7 @@ title: Language modeling a billion words
comments: True
author: nicholas-leonard
excerpt: Noise contrastive estimation is used to train a multi-GPU recurrent neural network language model on the Google billion words dataset.
-picture: https://raw.githubusercontent.com/torch/torch.github.io/master/blog/_posts/images/output_52iFki.gif
+picture: https://raw.githubusercontent.com/torch/torch.github.io/master/blog/_posts/images/rnnlm.png
---
<!---# Language modeling a billion words -->
@@ -911,9 +911,9 @@ Training is still underway, but to pique your interest, this is an example of ge
</SUBMISSION>
```
-This particular sample is a little depressing, but that might just be the nature of the `offmychest` subreddit...
-Conditioned on the opening `<SUBMISSION>` token, this generated sequence is incredibly human.
-Reading through the comment, I feel like I am reading a story written by an actual person.
+This particular sample is a little depressing, but that might just be the nature of the `offmychest` subreddit.
+Conditioned on the opening `<SUBMISSION>` token, this generated sequence, although imperfect, is incredibly human.
+Reading through the comment, I feel like I am reading a story written by an actual (somewhat schizophrenic) person.
The ability to similuate human creativity is one of the reasons I am so interested in using reddit data for language modeling.
A less depressing sample is the following, which concerns the [Destiny](https://en.wikipedia.org/wiki/Destiny_(video_game)) video game: