I published a guide to Keras preprocessing layers -- key new feature of the TensorFlow 2.3 release https://keras.io/guides/preprocessing_layers/ …pic.twitter.com/CvQoHYf9rS
Show this thread
-
-
Here, the hashing trick is effective; elsewhere it is being commonly used for word embeddings to disastrous effect, e.g. see https://towardsdatascience.com/the-hidden-costs-of-low-quality-word-embeddings-b05f38c7e095 … (pardon the self reference).
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Good trick
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What is a practical use case for this? E.g., if instead of the ints we hashed Tweet IDs into buckets, it is conceivable that we might hash a tweet by Trump and one by Bernie into the same bucket, causing them to get the same one-hot encoding.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.