Failed to import tinychat.

Installation is fine. The following is my setup. If I run the model_worker, error message with Failed to import tinychat.

PS: I am using the Qwen AWQ model. Qwen 14B AWQ

awq                           0.1.0        /home/user/tmp/FastChat/repositories/llm-awq
awq-inference-engine          0.0.0
fschat                        0.2.36       /home/user/tmp/FastChat
torch                         2.1.2+cu121
torchaudio                    2.1.2+cu121
torchvision                   0.16.2+cu121

python3 -m fastchat.serve.model_worker --model-name Qwen --model-path ./ --num-gpus 3 --awq-wbits 4 --awq-groupsize 128

2024-02-12 22:45:04 | INFO | model_worker | Loading the model ['Qwen'] on worker 2ed2e3b9 ...
2024-02-12 22:45:04 | INFO | stdout | Loading AWQ quantized model...
2024-02-12 22:45:04 | INFO | numexpr.utils | Note: NumExpr detected 48 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-12 22:45:04 | INFO | numexpr.utils | NumExpr defaulting to 8 threads.
2024-02-12 22:45:04 | INFO | datasets | PyTorch version 2.1.2+cu121 available.
2024-02-12 22:45:04 | INFO | stdout | Error: Failed to import tinychat. cannot import name 'real_quantize_model_weight' from 'awq.quantize.quantizer' (/home/chaoyou/anaconda3/envs/fastchat/lib/python3.10/site-packages/awq/quantize/quantizer.py)
2024-02-12 22:45:04 | INFO | stdout | Please double check if you have successfully installed AWQ
2024-02-12 22:45:04 | INFO | stdout | See https://github.com/lm-sys/FastChat/blob/main/docs/awq.md

Start at the fastchat.serve.model_worker entry point and review the AWQ setup described in docs/awq.md. Reproduce the Qwen AWQ command and trace the failed tinychat import; done means the model_worker starts without the reported import error.