Menu
I am trying to run Elasticsearch BERT application and would like to understand the minimal configuration for fine-tuning the model using GPU. What machine configuration should I be using?
Reference github: Fast-Bert
936
Compute-optimized - These machines provide the highest performance per core on Compute Engine and are optimized for compute-intensive workloads, such as high performance computing (HPC), game servers, and latency-sensitive API serving.
You can only attach GPUs to VMs using the N1 machine series or the A2 machine series. GPUs are not supported by other machine series. VMs with lower numbers of GPUs are limited to a maximum number of vCPUs. In general, a higher number of GPUs lets you create instances with a higher number of vCPUs and memory.
Before you create an instance with a GPU, select which boot disk image you want to use for the instance, and ensure that the appropriate GPU driver is installed. To create an instance with one or more GPUs using the Google Cloud Platform Console, Go to the VM instances page.
You would probably need to attach different GPUs to your compute instance to test performance. The Tesla T4 is the cheapest, while the Tesla V100 is the most expensive.
The n1-highmem or the n1-highcpu families of compute instance would be a good place to start.
Some of the specs published by Google:
64
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With
