Windows 11: Can't run LLama

[denmuslimov]

I'm experimenting with different models.
Some refuse to run.

Windows 11, RTX 4070 12Gb

The following commands:
python generate.py --base_model=TheBloke/Llama-2-7b-Chat-GPTQ --load_gptq="model" --use_safetensors=True --prompt_type=llama2 --save_dir='save'
python generate.py --base_model=TheBloke/Nous-Hermes-13B-GPTQ --score_model=None --load_gptq=model --use_safetensors=True --prompt_type=instruct --langchain_mode=UserData

Return the this error:

...
  File "C:\Users\Worker\Documents\AI\h2ogpt\generate.py", line 20, in <module>
    entrypoint_main()
  File "C:\Users\Worker\Documents\AI\h2ogpt\generate.py", line 16, in entrypoint_main
    H2O_Fire(main)
  File "C:\Users\Worker\Documents\AI\h2ogpt\src\utils.py", line 79, in H2O_Fire
    fire.Fire(component=component, command=args)
  File "C:\Users\Worker\miniconda3\envs\h2ogpt\lib\site-packages\fire\core.py", line 135, in Fire
    component_trace = _Fire(component, args, parsed_flag_args, context, name)
  File "C:\Users\Worker\miniconda3\envs\h2ogpt\lib\site-packages\fire\core.py", line 468, in _Fire
    component, remaining_args = _CallAndUpdateTrace(
  File "C:\Users\Worker\miniconda3\envs\h2ogpt\lib\site-packages\fire\core.py", line 684, in _CallAndUpdateTrace
    component = fn(*varargs, **kwargs)
  File "C:\Users\Worker\Documents\AI\h2ogpt\src\gen.py", line 2329, in main
    model_state_trial = model_lock_to_state(model_dict, cache_model_state=False, **kwargs_model_lock_to_state)
  File "C:\Users\Worker\Documents\AI\h2ogpt\src\model_utils.py", line 1689, in model_lock_to_state
    return __model_lock_to_state(model_dict1, **kwargs)
  File "C:\Users\Worker\Documents\AI\h2ogpt\src\model_utils.py", line 1767, in __model_lock_to_state
    model0, tokenizer0, device = get_model_retry(reward_type=False,
  File "C:\Users\Worker\Documents\AI\h2ogpt\src\model_utils.py", line 425, in get_model_retry
    model1, tokenizer1, device1 = get_model(**kwargs)
  File "C:\Users\Worker\Documents\AI\h2ogpt\src\model_utils.py", line 1208, in get_model
    return get_hf_model(load_8bit=load_8bit,
  File "C:\Users\Worker\Documents\AI\h2ogpt\src\model_utils.py", line 1534, in get_hf_model
    model = exllama_set_max_input_length(model, tokenizer.model_max_length)
  File "C:\Users\Worker\miniconda3\envs\h2ogpt\lib\site-packages\auto_gptq\utils\exllama_utils.py", line 15, in exllama_set_max_input_length
    from exllama_kernels import prepare_buffers, cleanup_buffers_cuda
ImportError: DLL load failed while importing exllama_kernels: The specified module could not be found.