You set `add_prefix_space`. The tokenizer needs to be converted from the slow tokenizers
#215
by d-hari - opened
You set
add_prefix_space. The tokenizer needs to be converted from the slow tokenizers
I get this message when loading the pipeline. It seems to be because add_prefix_space is set to true in the json file tokenizer_3/tokenizer_config.json, so I assume there is little I can do about it (I don't want to go messing with the config unless I know what I am doing).
Can I ignore this message, or is it something that Stability AI need to fix?
using gpu acceleration instead of cpu.
pipe = pipe.to("cuda")
It seems to be running quite fast. Can i comment this line out or is it something that SAI need to fix?
Hmm, I'm pretty sure I was doing that. But now when I try I do not get that warning message. Maybe it was because I was previously doing pipeline.enable_model_cpu_offload.
d-hari changed discussion status to closed