shape is invalid for input

#56
by Rwbyist - opened

model : gemma-7b-it
transformers:4.38.0

image.png
The above is my code. When I run generate, the following error occurred
RuntimeError: shape '[1, 2, 3072]' is invalid for input of size 8192

Running '2b-it' with the same code will not cause any issues.

deleted

simply upgrade the transformers model to 4.38.1

Google org

Closing this as it seems fixed, thanks @Saurabh19Mishra!

suryabhupa changed discussion status to closed

Sign up or log in to comment