Language Tuning Model - RESOURCE EXHAUSTED message

[nsubbia]

While tuning language model, the process stops every time at llm model tuning step with the below error  code=resource_exhausted, message=the following quota metrics exceed quota limits:"

When the Quota status is checked, the assigned quota limit for restricted_image_training_tpu_v3_pod  is 0. How to increase this value?

[kvandres]

Good day @nsubbia,

You need to file a request to get your quota increased, here is a link that you can use as a guide on how to file a request: https://cloud.google.com/docs/quota_detail/view_manage#requesting_higher_quota
Please note that it may be subject to approval, and this will be processed typically around 2-3 business days after that you will receive an email if the quota request was approved or denied. Also please note that some types of TPU v3 may not be available in some regions. For example, in us-central the only available type is TPU v3-8 while in the asia pacific, TPU v3 is not available. The region that contains the most types of TPU v3 is europe west. For more information about the availability of each TPU by region, you can visit this link: https://cloud.google.com/tpu/docs/regions-zones#overview

Hope this helps!