Someet

Hey @alvarobartt Thank you for the quick reply, I'll definitely check out the links.
Can you help me with one more think, So i have a use case where i have to classify like large data points like 5000-20000 into labels which are like 20-80 in like max 30-40 seconds, currently i use the zero shot classifier for them, using hugging face, but the inferencing are very slow. so i wanted to convert them to this endpoint so that i can give multiple request to this endpoint and get the job done.

i also tried using some 1b parameter llms using vllm on gpus for inferencing but they were not accurate.

can you help me out with a solution, what can be done in this case.

Hey, I'm getting error when i tried to do prediction on his zero shot classifier, i followed each step as it is given and when in the last step i tried to do prediction.
error: InternalServerError: 500 {"detail":"The following exception has occurred: TypeError. Arguments: ("ZeroShotClassificationPipeline.call() missing 1 required positional argument: 'sequences'",)."}

Can you help me identify why is this the case.