Got error when exporting model with quantization

[dickens88]

Reminder

• I have read the README and searched the existing issues.

Reproduction

Dears,

I'm using the latest code from master and i deploy my env with docker-compose. When i try to export model with quantization, the backend gives errors like this

llama_factory  | importlib.metadata.PackageNotFoundError: No package metadata was found for The 'optimum>=1.16.0' distribution was not found and is required by this application. 
llama_factory  | To fix: pip install optimum>=1.16.0

after i installed all the necessary packages with this command:

pip install auto_gptq
pip install optimum
pip install -U accelerate bitsandbytes datasets peft transformers

I got another error:

llama_factory  | Traceback (most recent call last):
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/queueing.py", line 566, in process_events
llama_factory  |     response = await route_utils.call_process_api(
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/route_utils.py", line 261, in call_process_api
llama_factory  |     output = await app.get_blocks().process_api(
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/blocks.py", line 1788, in process_api
llama_factory  |     result = await self.call_function(
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/blocks.py", line 1352, in call_function
llama_factory  |     prediction = await utils.async_iteration(iterator)
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/utils.py", line 583, in async_iteration
llama_factory  |     return await iterator.__anext__()
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/utils.py", line 576, in __anext__
llama_factory  |     return await anyio.to_thread.run_sync(
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/anyio/to_thread.py", line 56, in run_sync
llama_factory  |     return await get_async_backend().run_sync_in_worker_thread(
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/anyio/_backends/_asyncio.py", line 2144, in run_sync_in_worker_thread
llama_factory  |     return await future
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/anyio/_backends/_asyncio.py", line 851, in run
llama_factory  |     result = context.run(func, *args)
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/utils.py", line 559, in run_sync_iterator_async
llama_factory  |     return next(iterator)
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/gradio/utils.py", line 742, in gen_wrapper
llama_factory  |     response = next(iterator)
llama_factory  |   File "/app/src/llmtuner/webui/components/export.py", line 79, in save_model
llama_factory  |     export_model(args)
llama_factory  |   File "/app/src/llmtuner/train/tuner.py", line 57, in export_model
llama_factory  |     model = load_model(tokenizer, model_args, finetuning_args)  # must after fixing tokenizer to resize vocab
llama_factory  |   File "/app/src/llmtuner/model/loader.py", line 113, in load_model
llama_factory  |     model = AutoModelForCausalLM.from_pretrained(**init_kwargs)
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/transformers/models/auto/auto_factory.py", line 563, in from_pretrained
llama_factory  |     return model_class.from_pretrained(
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py", line 3165, in from_pretrained
llama_factory  |     hf_quantizer.validate_environment(
llama_factory  |   File "/usr/local/lib/python3.10/dist-packages/transformers/quantizers/quantizer_gptq.py", line 56, in validate_environment
llama_factory  |     raise ImportError(
llama_factory  | ImportError: Loading a GPTQ quantized model requires optimum (`pip install optimum`) and auto-gptq library (`pip install auto-gptq`)

my settings is like following:

Expected behavior

No response

System Info

No response

Others

No response