New discussion

Sequence length?

#5 opened over 1 year ago by deleted

Handle model parallelism

#4 opened over 1 year ago by
sgugger

adds _no_split_block

1
#3 opened almost 2 years ago by
staturecrane

Results are extremely poor

2
#2 opened almost 2 years ago by
Steve72

Quantization support.

3
#1 opened almost 2 years ago by
AV99