testing outlines suports with ROCM

[wizd]

requirements-rocm.txt missing outlines:

outlines >= 0.0.27

after adding outlines, running openai api server has error:

root@dualamd:/app# python -m vllm.entrypoints.openai.api_server \
> --gpu-memory-utilization 0.98 --trust-remote-code \
>     --served-model-name gpt-3.5-turbo-1106 \
>     --max-model-len 32768 --model Qwen1.5-14B
Traceback (most recent call last):
  File "/opt/conda/envs/py_3.9/lib/python3.9/runpy.py", line 197, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/opt/conda/envs/py_3.9/lib/python3.9/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/vllm-0.3.3+rocm603-py3.9-linux-x86_64.egg/vllm/entrypoints/openai/api_server.py", line 23, in <module>
    from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/vllm-0.3.3+rocm603-py3.9-linux-x86_64.egg/vllm/entrypoints/openai/serving_chat.py", line 15, in <module>
    from vllm.model_executor.guided_decoding import get_guided_decoding_logits_processor
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/vllm-0.3.3+rocm603-py3.9-linux-x86_64.egg/vllm/model_executor/guided_decoding.py", line 12, in <module>
    from vllm.model_executor.guided_logits_processors import JSONLogitsProcessor, RegexLogitsProcessor
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/vllm-0.3.3+rocm603-py3.9-linux-x86_64.egg/vllm/model_executor/guided_logits_processors.py", line 23, in <module>
    from outlines.fsm.fsm import RegexFSM
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/outlines/__init__.py", line 2, in <module>
    import outlines.generate
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/outlines/generate/__init__.py", line 1, in <module>
    from .api import SequenceGenerator
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/outlines/generate/api.py", line 5, in <module>
    from outlines.fsm.fsm import FSMState
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/outlines/fsm/fsm.py", line 9, in <module>
    from outlines.fsm.regex import create_fsm_index_tokenizer, make_deterministic_fsm
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/outlines/fsm/regex.py", line 5, in <module>
    import numba
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/numba/__init__.py", line 43, in <module>
    from numba.np.ufunc import (vectorize, guvectorize, threading_layer,
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/numba/np/ufunc/__init__.py", line 3, in <module>
    from numba.np.ufunc.decorators import Vectorize, GUVectorize, vectorize, guvectorize
  File "/opt/conda/envs/py_3.9/lib/python3.9/site-packages/numba/np/ufunc/decorators.py", line 3, in <module>
    from numba.np.ufunc import _internal
SystemError: initialization of _internal failed without raising an exception

it seems numba only supports CUDA but not ROCM.