PromtEngineer
Terminal VS Interface GPU problem
Hi all you guys,
My situation is that when I run the code from the terminal running on the GPU, it Goes very well on the GPU and very fast. BUT, when I run it through the API (with the same constant configuration) in the 90% of the cases it run with a low usage of the GPU (90% with terminal vs 6% with UI) and then in a very slow way.
I cannot control it. Could you help me please?
Thanks
Hi all you guys,
My situation is that when I run the code from the terminal running on the GPU, it Goes very well on the GPU and very fast. BUT, when I run it through the API (with the same constant configuration) in the 90% of the cases it run with a low usage of the GPU (90% with terminal vs 6% with UI) and then in a very slow way.
I cannot control it. Could you help me please?
Thanks
Hi, In my case GPU is only works during ingest process, rest all time (either its terminal or UI), its use CPU and Memory only. It will be great if you guide me to set specific parameter in respective file which help me to utilize gpu (24GB :-( ).
Hi, have you satisfied all the requirements asked in the documentation?
What kind of hardware do you have?
Hello,
Sorry, I don't have this issue and probably unintentially click the issue while browsing the github repository So far, I have been able to install localGPT and make it learn on of my document. Works fine. Some depedencies problem during the installation but linked to my environment more than the documentation. using mac studio M2 32gb. A bit slow but I'll tweak it.
Thanks for the job:)
Sent with Tuta; enjoy secure & ad-free emails: https://tuta.com
17 Jan 2024, 12:35 by @.***:
Hi, have you satisfied all the requirements asked in the documentation?
What kind of hardware do you have?
— Reply to this email directly, > view it on GitHub https://github.com/PromtEngineer/localGPT/issues/693#issuecomment-1895627349> , or > unsubscribe https://github.com/notifications/unsubscribe-auth/BCX4VW45XY45I2A2KGEWVS3YO6ZQLAVCNFSM6AAAAABA4NQB6SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOJVGYZDOMZUHE> . You are receiving this because you are subscribed to this thread.> Message ID: > <PromtEngineer/localGPT/issues/693/1895627349> @> github> .> com>
I got the same issue...no matter how, it always using cpu. Same, no matter what it is always using CPU. I think there is some problem with the usage of Torch and CUDA.
Any engineer here, master in these frameworks can do a PR for the following and it may work wonderfully.
Thank you. Nitkarsh Chourasia.
