Hui Chen comments

Repositories
Issues
Comments

Results 4 comments of


                                            Hui Chen

[BUG] Tensor are not on the same device when enable cpu activation offload

I encountered the same problem on deepspeed version 0.14.2.

batch_decode: TypeError: int() argument must be a string, a bytes-like object or a number, not 'list'

I also encountered this issue. Have you found any solutions?

batch_decode: TypeError: int() argument must be a string, a bytes-like object or a number, not 'list'

Have you tried to add ```predictions = np.argmax(predictions, axis=-1)``` before decoding? Current prediction shape looks like (batch_size, length, vocabulary_size). We should require a shape like (batch_size, length).

llama7B issue

Hi @JiuhaiChen , I also encountered this question. Did you solve it?