Partage de technologie

Analyse des outils open source pour convertir des PDF en Markdown

2024-07-12

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Marker : Analyse des outils open source pour convertir des PDF en Markdown
Marker est un projet open source développé sur GitHub par Vik Paruchuri. Sa fonction principale est de convertir des fichiers PDF au format Markdown. Voici une analyse détaillée du projet Marker :

Aperçu du projet:

Lien du projet : https://github.com/VikParuchuri/marker.git
Entretenu par : VikParuchuri
Principales fonctionnalités : convertissez rapidement et précisément le format PDF au format Markdown, prenant en charge plusieurs types de documents, en particulier les livres et les articles scientifiques.

Caractéristiques techniques:

Modèle d'apprentissage profond : Marker utilise une série de modèles d'apprentissage profond pour extraire du texte, détecter la mise en page, nettoyer et formater les blocs de texte, et enfin les combiner dans des documents Markdown.
Prise en charge de l'OCR : pour les scénarios nécessitant l'OCR, Marker prend en charge l'utilisation d'outils OCR tels que Surya et Tesseract pour garantir l'exactitude de l'extraction de texte.
Prise en charge multiplateforme : Marker peut fonctionner sur GPU, CPU ou MPS pour répondre aux besoins de différents environnements matériels.

Détails des fonctionnalités :

Traitement des documents : prend en charge la suppression des en-têtes, des pieds de page et autres impuretés, le formatage des tableaux et des blocs de code, l'extraction et l'enregistrement des images.
Prise en charge des langues : Marker prend en charge toutes les langues et les utilisateurs peuvent optimiser l'effet OCR en spécifiant une liste de langues.
Conversion d'équations : capable de convertir la plupart des équations au format LaTeX, ce qui facilite l'intégration de formules mathématiques dans les documents Markdown.

Performance:

Vitesse et précision : Marker excelle en vitesse et en précision, ce qui lui confère des avantages significatifs, notamment par rapport à d’autres outils tels que le nougat.
Utilisation des ressources : lors de l'exécution sur A6000 Ada, chaque tâche occupe en moyenne environ 4 Go de VRAM, prenant en charge le traitement parallèle de plusieurs documents.

Conseils d'utilisation :

Installation : les utilisateurs doivent installer le package Marker-pdf via pip

pip install marker-pdf 


  • 1
  • 2
(GraphRAG) PS D:python-workspaceGraphRAG> pip install marker-pdf 
Looking in indexes: https://mirrors.aliyun.com/pypi/simple/
Collecting marker-pdf
  Downloading https://mirrors.aliyun.com/pypi/packages/05/c1/782f56407ea60bd35c127c829b8e43da99a0da41f6c9ee002cab97e430c5/marker_pdf-0.2.15-py3-none-any.whl (63 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 63.2/63.2 kB 563.9 kB/s eta 0:00:00
Requirement already satisfied: Pillow<11.0.0,>=10.1.0 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (10.4.0)
Requirement already satisfied: filetype<2.0.0,>=1.2.0 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (1.2.0)
Collecting ftfy<7.0.0,>=6.1.1 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/f4/f0/21efef51304172736b823689aaf82f33dbc64f54e9b046b75f5212d5cee7/ftfy-6.2.0-py3-none-any.whl (54 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 54.4/54.4 kB 353.5 kB/s eta 0:00:00
Requirement already satisfied: grpcio<2.0.0,>=1.63.0 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (1.64.1)
Requirement already satisfied: numpy<2.0.0,>=1.26.1 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (1.26.4)
Collecting pdftext<0.4.0,>=0.3.10 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/54/78/8dd39d5ed3b90fb7ecaa20f92ff09c4594877a88501de6352d22e8c53aa0/pdftext-0.3.10-py3-none-any.whl (25 kB)
Requirement already satisfied: pydantic<3.0.0,>=2.4.2 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (2.8.0)
Collecting pydantic-settings<3.0.0,>=2.0.3 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/e8/4f/aad03d5f711717d94d7de9684cb542343b392df1ad6889118636674fc983/pydantic_settings-2.3.4-py3-none-any.whl (22 kB)
Requirement already satisfied: python-dotenv<2.0.0,>=1.0.0 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (1.0.1)
Collecting rapidfuzz<4.0.0,>=3.8.1 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/60/a6/6c2f5e9be933150a6d55ffce4ff6d9701ddfc5b267c789a84674eadbd373/rapidfuzz-3.9.4-cp311-cp311-win_amd64.whl (1.6 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.6/1.6 MB 684.6 kB/s eta 0:00:00
Requirement already satisfied: regex<2025.0.0,>=2024.4.28 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (2024.5.15)
Requirement already satisfied: scikit-learn<2.0.0,>=1.3.2 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (1.5.0)
Collecting surya-ocr<0.5.0,>=0.4.14 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/62/a8/dd78c484fa9a459e388a31aa3a45d23eb454c6aeb2a17710284631088615/surya_ocr-0.4.14-py3-none-any.whl (94 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 94.5/94.5 kB 773.2 kB/s eta 0:00:00
Collecting tabulate<0.10.0,>=0.9.0 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/40/44/4a5f08c96eb108af5cb50b41f76142f0afa346dfa99d5296fe7202a11854/tabulate-0.9.0-py3-none-any.whl (35 kB)
Collecting texify<0.2.0,>=0.1.10 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/76/26/c12d194dd90bd78b524a7054e9125685efc32149d29005ca61c72ff4c126/texify-0.1.10-py3-none-any.whl (30 kB)
Collecting torch<3.0.0,>=2.2.2 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/d3/1d/a257913c89572de61316461db91867f87519146e58132cdeace3d9ffbe1f/torch-2.3.1-cp311-cp311-win_amd64.whl (159.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 159.8/159.8 MB 635.4 kB/s eta 0:00:00
Requirement already satisfied: tqdm<5.0.0,>=4.66.1 in e:programdataminiconda3envsgraphraglibsite-packages (from marker-pdf) (4.66.4)
Collecting transformers<5.0.0,>=4.36.2 (from marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/20/5c/244db59e074e80248fdfa60495eeee257e4d97c3df3487df68be30cd60c8/transformers-4.42.3-py3-none-any.whl (9.3 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 9.3/9.3 MB 639.8 kB/s eta 0:00:00
Collecting wcwidth<0.3.0,>=0.2.12 (from ftfy<7.0.0,>=6.1.1->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/fd/84/fd2ba7aafacbad3c4201d395674fc6348826569da3c0937e75505ead3528/wcwidth-0.2.13-py2.py3-none-any.whl (34 kB)
Collecting pypdfium2<5.0.0,>=4.29.0 (from pdftext<0.4.0,>=0.3.10->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/25/bd/56d9ec6b9f0fc4e0d95288759f3179f0fcd34b1a1526b75673d2f6d5196f/pypdfium2-4.30.0-py3-none-win_amd64.whl (2.9 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.9/2.9 MB 627.3 kB/s eta 0:00:00
Requirement already satisfied: annotated-types>=0.4.0 in e:programdataminiconda3envsgraphraglibsite-packages (from pydantic<3.0.0,>=2.4.2->marker-pdf) (0.7.0)
Requirement already satisfied: pydantic-core==2.20.0 in e:programdataminiconda3envsgraphraglibsite-packages (from pydantic<3.0.0,>=2.4.2->marker-pdf) (2.20.0)
Requirement already satisfied: typing-extensions>=4.6.1 in e:programdataminiconda3envsgraphraglibsite-packages (from pydantic<3.0.0,>=2.4.2->marker-pdf) (4.12.2)
Requirement already satisfied: scipy>=1.6.0 in e:programdataminiconda3envsgraphraglibsite-packages (from scikit-learn<2.0.0,>=1.3.2->marker-pdf) (1.12.0)
Requirement already satisfied: joblib>=1.2.0 in e:programdataminiconda3envsgraphraglibsite-packages (from scikit-learn<2.0.0,>=1.3.2->marker-pdf) (1.4.2)
Requirement already satisfied: threadpoolctl>=3.1.0 in e:programdataminiconda3envsgraphraglibsite-packages (from scikit-learn<2.0.0,>=1.3.2->marker-pdf) (3.5.0)
Collecting opencv-python<5.0.0.0,>=4.9.0.80 (from surya-ocr<0.5.0,>=0.4.14->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/ec/6c/fab8113424af5049f85717e8e527ca3773299a3c6b02506e66436e19874f/opencv_python-4.10.0.84-cp37-abi3-win_amd64.whl (38.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 38.8/38.8 MB 546.5 kB/s eta 0:00:00
Collecting filelock (from torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/ae/f0/48285f0262fe47103a4a45972ed2f9b93e4c80b8fd609fa98da78b2a5706/filelock-3.15.4-py3-none-any.whl (16 kB)
Collecting sympy (from torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/61/53/e18c8c97d0b2724d85c9830477e3ebea3acf1dcdc6deb344d5d9c93a9946/sympy-1.12.1-py3-none-any.whl (5.7 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 5.7/5.7 MB 624.0 kB/s eta 0:00:00
Requirement already satisfied: networkx in e:programdataminiconda3envsgraphraglibsite-packages (from torch<3.0.0,>=2.2.2->marker-pdf) (3.3)
Collecting jinja2 (from torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/31/80/3a54838c3fb461f6fec263ebf3a3a41771bd05190238de3486aae8540c36/jinja2-3.1.4-py3-none-any.whl (133 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 133.3/133.3 kB 492.1 kB/s eta 0:00:00
Requirement already satisfied: fsspec in e:programdataminiconda3envsgraphraglibsite-packages (from torch<3.0.0,>=2.2.2->marker-pdf) (2024.6.1)
Collecting mkl<=2021.4.0,>=2021.1.1 (from torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/fe/1c/5f6dbf18e8b73e0a5472466f0ea8d48ce9efae39bd2ff38cebf8dce61259/mkl-2021.4.0-py2.py3-none-win_amd64.whl (228.5 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 228.5/228.5 MB 597.0 kB/s eta 0:00:00
Requirement already satisfied: colorama in e:programdataminiconda3envsgraphraglibsite-packages (from tqdm<5.0.0,>=4.66.1->marker-pdf) (0.4.6)
Collecting huggingface-hub<1.0,>=0.23.2 (from transformers<5.0.0,>=4.36.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/69/d6/73f9d1b7c4da5f0544bc17680d0fa9932445423b90cd38e1ee77d001a4f5/huggingface_hub-0.23.4-py3-none-any.whl (402 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 402.6/402.6 kB 598.2 kB/s eta 0:00:00
Requirement already satisfied: packaging>=20.0 in e:programdataminiconda3envsgraphraglibsite-packages (from transformers<5.0.0,>=4.36.2->marker-pdf) (23.2)
Requirement already satisfied: pyyaml>=5.1 in e:programdataminiconda3envsgraphraglibsite-packages (from transformers<5.0.0,>=4.36.2->marker-pdf) (6.0.1)
Requirement already satisfied: requests in e:programdataminiconda3envsgraphraglibsite-packages (from transformers<5.0.0,>=4.36.2->marker-pdf) (2.32.3)
Collecting safetensors>=0.4.1 (from transformers<5.0.0,>=4.36.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/cb/f6/19f268662be898ff2a23ac06f8dd0d2956b2ecd204c96e1ee07ba292c119/safetensors-0.4.3-cp311-none-win_amd64.whl (287 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 287.3/287.3 kB 571.3 kB/s eta 0:00:00
Collecting tokenizers<0.20,>=0.19 (from transformers<5.0.0,>=4.36.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/65/8e/6d7d72b28f22c422cff8beae10ac3c2e4376b9be721ef8167b7eecd1da62/tokenizers-0.19.1-cp311-none-win_amd64.whl (2.2 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.2/2.2 MB 625.9 kB/s eta 0:00:00
Collecting intel-openmp==2021.* (from mkl<=2021.4.0,>=2021.1.1->torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/6f/21/b590c0cc3888b24f2ac9898c41d852d7454a1695fbad34bee85dba6dc408/intel_openmp-2021.4.0-py2.py3-none-win_amd64.whl (3.5 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 3.5/3.5 MB 487.9 kB/s eta 0:00:00
Collecting tbb==2021.* (from mkl<=2021.4.0,>=2021.1.1->torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/f1/24/500811330b3b070e5995c3275181dbcd00c06cef26c6ebfe6ee1ca9b6223/tbb-2021.13.0-py3-none-win_amd64.whl (286 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 286.9/286.9 kB 505.8 kB/s eta 0:00:00
Collecting MarkupSafe>=2.0 (from jinja2->torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/b7/a2/c78a06a9ec6d04b3445a949615c4c7ed86a0b2eb68e44e7541b9d57067cc/MarkupSafe-2.1.5-cp311-cp311-win_amd64.whl (17 kB)
Requirement already satisfied: charset-normalizer<4,>=2 in e:programdataminiconda3envsgraphraglibsite-packages (from requests->transformers<5.0.0,>=4.36.2->marker-pdf) (3.3.2)
Requirement already satisfied: idna<4,>=2.5 in e:programdataminiconda3envsgraphraglibsite-packages (from requests->transformers<5.0.0,>=4.36.2->marker-pdf) (3.7)
Requirement already satisfied: urllib3<3,>=1.21.1 in e:programdataminiconda3envsgraphraglibsite-packages (from requests->transformers<5.0.0,>=4.36.2->marker-pdf) (2.2.2)
Requirement already satisfied: certifi>=2017.4.17 in e:programdataminiconda3envsgraphraglibsite-packages (from requests->transformers<5.0.0,>=4.36.2->marker-pdf) (2024.6.2)
Collecting mpmath<1.4.0,>=1.1.0 (from sympy->torch<3.0.0,>=2.2.2->marker-pdf)
  Downloading https://mirrors.aliyun.com/pypi/packages/43/e3/7d92a15f894aa0c9c4b49b8ee9ac9850d6e63b03c9c32c0367a13ae62209/mpmath-1.3.0-py3-none-any.whl (536 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 536.2/536.2 kB 495.3 kB/s eta 0:00:00
Installing collected packages: wcwidth, tbb, mpmath, intel-openmp, tabulate, sympy, safetensors, rapidfuzz, pypdfium2
, opencv-python, mkl, MarkupSafe, ftfy, filelock, jinja2, huggingface-hub, torch, tokenizers, pydantic-settings, transformers, pdftext, texify, surya-ocr, marker-pdf
Successfully installed MarkupSafe-2.1.5 filelock-3.15.4 ftfy-6.2.0 huggingface-hub-0.23.4 intel-openmp-2021.4.0 jinja2-3.1.4 marker-pdf-0.2.15 mkl-2021.4.0 mpmath-1.3.0 opencv-python-4.10.0.84 pdftext-0.3.10 pydantic-settings-2.3.4 pypdfium2-4.30.0 rapidfuzz-3.9.4 safetensors-0.4.3 surya-ocr-0.4.14 sympy-1.12.1 tabulate-0.9.0 tbb-2021.13.0 texify-0.1.10 tokenizers-0.19.1 torch-2.3.1 transformers-4.42.3 wcwidth-0.2.13


使用示例：

```bash
(GraphRAG) PS D:python-workspaceGraphRAG> marker_single GPT.pdf ./folder --batch_multiplier 2 --max_pages 52 --langs English
config.json: 100%|█████████████████████████████████████████████████████████████████████| 1.18k/1.18k [00:00<?, ?B/s] 
model.safetensors: 100%|█████████████████████████████████████████████████████████| 120M/120M [00:07<00:00, 16.7MB/s] 
Loaded detection model vikp/surya_det2 on device cpu with dtype torch.float32
preprocessor_config.json: 100%|████████████████████████████████████████████████████████████| 430/430 [00:00<?, ?B/s] 
config.json: 100%|█████████████████████████████████████████████████████████████████████| 1.57k/1.57k [00:00<?, ?B/s] 
model.safetensors: 100%|█████████████████████████████████████████████████████████| 120M/120M [00:06<00:00, 18.0MB/s] 
Loaded detection model vikp/surya_layout2 on device cpu with dtype torch.float32
preprocessor_config.json: 100%|████████████████████████████████████████████████████████████| 430/430 [00:00<?, ?B/s] 
config.json: 100%|█████████████████████████████████████████████████████████████████████| 5.04k/5.04k [00:00<?, ?B/s] 
model.safetensors: 100%|█████████████████████████████████████████████████████████| 550M/550M [00:34<00:00, 16.2MB/s] 
generation_config.json: 100%|██████████████████████████████████████████████████████████████| 160/160 [00:00<?, ?B/s] 
Loaded reading order model vikp/surya_order on device cpu with dtype torch.float32
preprocessor_config.json: 100%|████████████████████████████████████████████████████████████| 684/684 [00:00<?, ?B/s] 
config.json: 100%|█████████████████████████████████████████████████████████████| 6.91k/6.91k [00:00<00:00, 6.82MB/s] 
model.safetensors: 100%|███████████████████████████████████████████████████████| 1.05G/1.05G [01:04<00:00, 16.2MB/s] 
generation_config.json: 100%|██████████████████████████████████████████████████████████████| 181/181 [00:00<?, ?B/s]
Loaded recognition model vikp/surya_rec on device cpu with dtype torch.float32
preprocessor_config.json: 100%|█████████████████████████████████████████████████████| 608/608 [00:00<00:00, 605kB/s]
config.json: 100%|█████████████████████████████████████████████████████████████████████| 4.92k/4.92k [00:00<?, ?B/s]
model.safetensors: 100%|█████████████████████████████████████████████████████████| 625M/625M [00:38<00:00, 16.4MB/s]
generation_config.json: 100%|██████████████████████████████████████████████████████████████| 191/191 [00:00<?, ?B/s]
Loaded texify model to cpu with torch.float32 dtype
preprocessor_config.json: 100%|████████████████████████████████████████████████████████████| 617/617 [00:00<?, ?B/s]
tokenizer_config.json: 100%|███████████████████████████████████████████████████████████| 4.49k/4.49k [00:00<?, ?B/s]
tokenizer.json: 100%|██████████████████████████████████████████████████████████| 2.14M/2.14M [00:00<00:00, 2.85MB/s]
added_tokens.json: 100%|███████████████████████████████████████████████████████████████| 18.3k/18.3k [00:00<?, ?B/s]
special_tokens_map.json: 100%|█████████████████████████████████████████████████████| 552/552 [00:00<00:00, 6.29MB/s] 
Detecting bboxes: 100%|███████████████████████████████████████████████████████████████| 7/7 [05:49<00:00, 49.99s/it] 
Recognizing Text: 100%|███████████████████████████████████████████████████████████████| 1/1 [00:11<00:00, 11.37s/it] 
Detecting bboxes: 100%|███████████████████████████████████████████████████████████████| 5/5 [05:32<00:00, 66.45s/it] 
Finding reading order: 100%|██████████████████████████████████████████████████████████| 5/5 [03:15<00:00, 39.04s/it] 
Saved markdown to the ./folderGPT folder


  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40
  • 41
  • 42
  • 43
  • 44
  • 45
  • 46
  • 47
  • 48
  • 49
  • 50
  • 51
  • 52
  • 53
  • 54
  • 55
  • 56
  • 57
  • 58
  • 59
  • 60
  • 61
  • 62
  • 63
  • 64
  • 65
  • 66
  • 67
  • 68
  • 69
  • 70
  • 71
  • 72
  • 73
  • 74
  • 75
  • 76
  • 77
  • 78
  • 79
  • 80
  • 81
  • 82
  • 83
  • 84
  • 85
  • 86
  • 87
  • 88
  • 89
  • 90
  • 91
  • 92
  • 93
  • 94
  • 95
  • 96
  • 97
  • 98
  • 99
  • 100
  • 101
  • 102
  • 103
  • 104
  • 105
  • 106
  • 107
  • 108
  • 109
  • 110
  • 111
  • 112
  • 113
  • 114
  • 115
  • 116
  • 117
  • 118
  • 119
  • 120
  • 121
  • 122
  • 123
  • 124
  • 125
  • 126
  • 127
  • 128
  • 129
  • 130
  • 131
  • 132
  • 133
  • 134

配置：用户可以通过环境变量或配置文件调整Marker的行为，如设置OCR引擎、指定GPU设备、配置内存使用等。
命令行工具：Marker提供了命令行工具，允许用户以批处理方式转换单个或多个PDF文件。




商业使用与许可：

商业限制：虽然研究和个人使用是免费的，但商业使用受到一定限制。模型权重采用cc-by-nc-sa-4.0许可证，但作者为符合条件的小型组织提供了许可证豁免。
双许可选项：对于需要去除GPL许可证要求或超出收入限制的商业用户，提供了双许可选项。


社区与支持：

Discord社区：用户可以在Discord上讨论Marker的未来开发和其他相关问题。
文档与示例：GitHub仓库提供了详细的文档和示例，帮助用户快速上手。



总结：
Marker是一个功能强大、易于使用的PDF转Markdown工具，通过深度学习模型和OCR技术的结合，实现了高效且准确的文档转换。它不仅支持多种文档类型和语言，还提供了丰富的配置选项和命令行工具，满足了不同用户的需求。同时，Marker的社区支持和文档也非常完善，为用户提供了良好的使用体验。

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22

Profil personnel

Il se consacre à la recherche technologique depuis plus de 30 ans et maîtrise divers langages tels que java, linux, javascript, php, css, etc. Il a apporté de nombreuses contributions dans le domaine de l'open source. station de documentation pour les développeurs pour partager certains problèmes de développement technologique pour référence future.

mes coordonnées

Mail