You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the sdxl jax blog, example inference in the blog code is given and blog says with that inference 4 images is generated about 2 s with Cloud TPU v5e-4. However, Even though I am using same code and google cloud same tpu with ubuntu 22.04 base software version and python 3.10, I am generating images 3.23 seconds. Could you help me about what am I missing to achieve fast inference?
In the sdxl jax blog, example inference in the blog code is given and blog says with that inference 4 images is generated about 2 s with Cloud TPU v5e-4. However, Even though I am using same code and google cloud same tpu with ubuntu 22.04 base software version and python 3.10, I am generating images 3.23 seconds. Could you help me about what am I missing to achieve fast inference?
The text was updated successfully, but these errors were encountered: