Object Detection
Transformers
Safetensors
rt_detr_v2
document-layout-analysis
finance
rtdetr-v2
cuda
tf32
ax-pdf
Instructions to use AutomatosX/AXDocLayout-M-v1.4-CUDA-TF32 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AutomatosX/AXDocLayout-M-v1.4-CUDA-TF32 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("object-detection", model="AutomatosX/AXDocLayout-M-v1.4-CUDA-TF32")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForObjectDetection tokenizer = AutoTokenizer.from_pretrained("AutomatosX/AXDocLayout-M-v1.4-CUDA-TF32") model = AutoModelForObjectDetection.from_pretrained("AutomatosX/AXDocLayout-M-v1.4-CUDA-TF32", device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 629 Bytes
d2446c1 | 1 2 3 4 5 6 7 8 9 10 11 | AX-DocLayout-M v1.4 (H5)
Fine-tuned model and inference example: AutomatosX,2026.
This model is derived from docling-project/docling-layout-heron,
revision8f39ad3c0b4c58e9c2d2c84a38465abf757272d8, distributed under Apache-2.0.
The upstream model card and the Docling authors' technical report provide the
original model attribution. See https://huggingface.co/docling-project/docling-layout-heron
and https://arxiv.org/abs/2509.11720 (Livathinos et al.,2025).
Changes: H1/H5 finance fine-tuning, AX PDF SDK class binding, and release packaging.
No upstream logo, training documents, PDF parsing code or corpus is redistributed.
|