Speed benchmark
#2
by
rom1504
- opened
Hey,
Cool model!
Could you please share a speed benchmark of this model vs dense B/32 ?
Doing it on GPU and on CPU (and sharing the specific hardware used) would be great
have install pip install deepsparse-nightly[clip]>=1.7.0.20231210
but got error:
ValueError: Unknown Pipeline task clip_visual. Currently supported tasks are ['haystack', 'zero_shot_text_classification', 'open_pif_paf', 'ner', 'yolov8', 'qa', 'chat', 'codegen', 'glue', 'embedding_extraction', 'transformers_embedding_extraction', 'text_generation', 'token_classification', 'opt', 'bloom', 'information_retrieval_haystack', 'image_classification', 'yolo', 'code_generation', 'question_answering', 'text_classification', 'sentiment_analysis', 'custom', 'yolact', 'chatbot']