50系显卡docker支持

[sysugxh]

请问50系显卡可以使用docker部署吗，测试了几次构建容器都失败了

[sysugxh]

已按照#90的方式修改了cuda 版本
PS C:\Users\x\Downloads\MonkeyOCR-main\MonkeyOCR-main\docker> docker compose up monkeyocr-api
Attaching to monkeyocr-api-1
monkeyocr-api-1 | [INFO] === MonkeyOCR Docker Container Starting ===
monkeyocr-api-1 | [INFO] GPU Information:
monkeyocr-api-1 | NVIDIA GeForce RTX 5090, 32607
monkeyocr-api-1 | [INFO] Checking and downloading models...
monkeyocr-api-1 | [INFO] All required model files are present.
monkeyocr-api-1 | [INFO] Model files already exist, skipping download
monkeyocr-api-1 | [INFO] Models ready
monkeyocr-api-1 | [INFO] Starting FastAPI server
monkeyocr-api-1 | /usr/local/lib/python3.10/dist-packages/torch/cuda/init.py:61: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
monkeyocr-api-1 | import pynvml # type: ignore[import]
monkeyocr-api-1 | 2025-10-20 13:33:31.643 | INFO | api.main::189 - Using temporary directory: /app/tmp
monkeyocr-api-1 | INFO: Started server process [1]
monkeyocr-api-1 | INFO: Waiting for application startup.
monkeyocr-api-1 | 2025-10-20 13:33:31.649 | INFO | magic_pdf.model.custom_model:init:30 - using configs: {'device': 'cuda', 'weights': {'doclayout_yolo': 'Structure/doclayout_yolo_docstructbench_imgsz1280_2501.pt', 'PP-DocLayout_plus-L': 'Structure/PP-DocLayout_plus-L', 'layoutreader': 'Relation'}, 'models_dir': 'model_weight', 'layout_config': {'model': 'PP-DocLayout_plus-L', 'reader': {'name': 'layoutreader'}}, 'chat_config': {'weight_path': 'model_weight/Recognition', 'backend': 'lmdeploy', 'data_parallelism': 1, 'model_parallelism': 1, 'batch_size': 10, 'queue_config': {'max_batch_size': 256, 'queue_timeout': 1, 'max_queue_size': 2000}}}
monkeyocr-api-1 | 2025-10-20 13:33:31.649 | INFO | magic_pdf.model.custom_model:init:33 - using device: cuda
monkeyocr-api-1 | /usr/local/lib/python3.10/dist-packages/torch/cuda/init.py:287: UserWarning:
monkeyocr-api-1 | NVIDIA GeForce RTX 5090 with CUDA capability sm_120 is not compatible with the current PyTorch installation.
monkeyocr-api-1 | The current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_70 sm_75 sm_80 sm_86 sm_90.
monkeyocr-api-1 | If you want to use the NVIDIA GeForce RTX 5090 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
monkeyocr-api-1 |
monkeyocr-api-1 | warnings.warn(
monkeyocr-api-1 | 2025-10-20 13:33:31.749 | INFO | magic_pdf.model.custom_model:init:45 - using models_dir: model_weight
monkeyocr-api-1 | 2025-10-20 13:33:36.349 | INFO | api.main:lifespan:171 - ❌ Failed to initialize MonkeyOCR model: Paddlex is not installed. Please install it using 'pip install paddlex'.
monkeyocr-api-1 | ERROR: Traceback (most recent call last):
monkeyocr-api-1 | File "/app/MonkeyOCR/magic_pdf/model/sub_modules/layout/paddlex_layout/PaddleXLayoutModel.py", line 9, in
monkeyocr-api-1 | from paddlex import create_model
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/init.py", line 49, in
monkeyocr-api-1 | from .inference import create_pipeline, create_predictor
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/init.py", line 17, in
monkeyocr-api-1 | from .pipelines import create_pipeline, load_pipeline_config
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/pipelines/init.py", line 23, in
monkeyocr-api-1 | from .attribute_recognition import (
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/pipelines/attribute_recognition/init.py", line 15, in
monkeyocr-api-1 | from .pipeline import PedestrianAttributeRecPipeline, VehicleAttributeRecPipeline
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/pipelines/attribute_recognition/pipeline.py", line 27, in
monkeyocr-api-1 | from ..components import CropByBoxes
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/pipelines/components/init.py", line 29, in
monkeyocr-api-1 | from .retriever.base import BaseRetriever
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/pipelines/components/retriever/init.py", line 15, in
monkeyocr-api-1 | from .openai_bot_retriever import OpenAIBotRetriever
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/pipelines/components/retriever/openai_bot_retriever.py", line 16, in
monkeyocr-api-1 | from .base import BaseRetriever
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/paddlex/inference/pipelines/components/retriever/base.py", line 25, in
monkeyocr-api-1 | from langchain.docstore.document import Document
monkeyocr-api-1 | ModuleNotFoundError: No module named 'langchain.docstore'
monkeyocr-api-1 |
monkeyocr-api-1 | During handling of the above exception, another exception occurred:
monkeyocr-api-1 |
monkeyocr-api-1 | Traceback (most recent call last):
monkeyocr-api-1 | File "/usr/local/lib/python3.10/dist-packages/starlette/routing.py", line 694, in lifespan
monkeyocr-api-1 | async with self.lifespan_context(app) as maybe_state:
monkeyocr-api-1 | File "/usr/lib/python3.10/contextlib.py", line 199, in aenter
monkeyocr-api-1 | return await anext(self.gen)
monkeyocr-api-1 | File "/app/MonkeyOCR/api/main.py", line 167, in lifespan
monkeyocr-api-1 | initialize_model()
monkeyocr-api-1 | File "/app/MonkeyOCR/api/main.py", line 53, in initialize_model
monkeyocr-api-1 | monkey_ocr_model = MonkeyOCR(config_path)
monkeyocr-api-1 | File "/app/MonkeyOCR/magic_pdf/model/custom_model.py", line 80, in init
monkeyocr-api-1 | self.layout_model = atom_model_manager.get_atom_model(
monkeyocr-api-1 | File "/app/MonkeyOCR/magic_pdf/model/sub_modules/model_init.py", line 43, in get_atom_model
monkeyocr-api-1 | self._models[key] = atom_model_init(model_name=atom_model_name, **kwargs)
monkeyocr-api-1 | File "/app/MonkeyOCR/magic_pdf/model/sub_modules/model_init.py", line 56, in atom_model_init
monkeyocr-api-1 | atom_model = paddex_layout_model_init(
monkeyocr-api-1 | File "/app/MonkeyOCR/magic_pdf/model/sub_modules/model_init.py", line 18, in paddex_layout_model_init
monkeyocr-api-1 | from magic_pdf.model.sub_modules.layout.paddlex_layout.PaddleXLayoutModel import
monkeyocr-api-1 | File "/app/MonkeyOCR/magic_pdf/model/sub_modules/layout/paddlex_layout/PaddleXLayoutModel.py", line 11, in
monkeyocr-api-1 | raise ImportError("Paddlex is not installed. Please install it using 'pip install paddlex'.")
monkeyocr-api-1 | ImportError: Paddlex is not installed. Please install it using 'pip install paddlex'.
monkeyocr-api-1 |
monkeyocr-api-1 | ERROR: Application startup failed. Exiting.
monkeyocr-api-1 exited with code 3
PS C:\Users\x\Downloads\MonkeyOCR-main\MonkeyOCR-main\docker> 提示这个

[zenosai]

你好，看起来是没有安装Paddlex， #90 中的回复是针对之前的MonkeyOCR版本的，现在pro版本还需要安装paddle，不知道是否仍然适配