transformers
f73a4db3 - Inclusion of process_group in the gather_full_tensor function in tensor_parallel.py (#43932)

Commit
145 days ago
Inclusion of process_group in the gather_full_tensor function in tensor_parallel.py (#43932) * Bug fix on gather_full_tensor for upstreaming. * Made correct usage of process_group in all_gather now.
Author
Parents
Loading