AutoModelForCausalLM.from_pretrained() error
#2
by
liyang31163150
- opened
torch.load() error
magic_number = pickle_module.load(f, **pickle_load_args)
_pickle.UnpicklingError: could not find MARK
Can U help me, please?
This isn't a model that works with torch. This is a ggml for llama.cpp (cpu inference)