root@a:~/qwen/qwen.cpp/qwen_cpp# python3 convert.py -i /root/qwen/Qwen1.5-1.8B -t q4_0 -o qwen1_8b.bin
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
Traceback (most recent call last):
File "/root/qwen/qwen.cpp/qwen_cpp/convert.py", line 262, in
main()
File "/root/qwen/qwen.cpp/qwen_cpp/convert.py", line 256, in main
convert(f, args.model_name_or_path, dtype=args.type)
File "/root/qwen/qwen.cpp/qwen_cpp/convert.py", line 230, in convert
QwenConverter.convert(f, model, tokenizer, ggml_type)
File "/root/qwen/qwen.cpp/qwen_cpp/convert.py", line 182, in convert
cls.dump_config(f, model.config, model.generation_config, tokenizer, ggml_type)
File "/root/qwen/qwen.cpp/qwen_cpp/convert.py", line 198, in dump_config
tokenizer.im_start_id,
AttributeError: 'Qwen2TokenizerFast' object has no attribute 'im_start_id'