Debug fine-tuning pipeline for GPT-J


I am trying to train the model on a couple hundred megabytes of fine-tuning data using v3-8 tpus on Google Cloud.

I got some error

RuntimeError: Internal: Replica group has size 8, but all replica groups in an all-to-all with N operands must have size N:

I need expert's help


Evner: Machine Learning (ML), Tensorflow, Deep Learning, Pytorch

Se mere: language fine tuning, fine tuning courier imap, fine tuning oracle application vsat, fine tuning adwords, adobe premiere fine tuning, website fine tuning, phplist fine tuning, oracle fine tuning vsat network, fine tuning javascript, fine tuning joomla, lamp fine tuning jobs, php fine tuning projects, postfix fine tuning postfix org, seo fine tuning, green screen fine tuning final cut, fine tuning effects, mysql fine tuning, fine tuning joomla website, asterisk fine tuning, gpt 2 fine tuning github

Om arbejdsgiveren:
( 51 bedømmelser ) Phnom Penh, Cambodia

Projekt ID: #30587875