LlavaNextProcessor from_pretrained Fails - tokenizer.json Loading Error
#2
by CrystalTang - opened
Exception at: model.set_processor("BAAI/BGE-VL-MLLM-S1")
When trying to load the BGE-VL-MLLM-S1 model, I encountered a tokenizer.json parsing error, even though the file was correctly downloaded. I also tried setting use_fast=False, but the issue persists.（my environment: transformers==4.44.2, Python 3.9.18, torch+CUDA: 2.0.1+cu117)