Dataset · Teklia
English handwritten text lines for OCR
Teklia/IAM-line
A classical dataset of individual handwritten text line images in English for OCR training.
What's inside
A version of the IAM Handwriting Database sliced at the individual line level (ranging from 10 to 100 thousand image-text pairs). All documents are written in English, and the images are normalized to a fixed height of 128 pixels.
What you can do with it
- Train and test optical character recognition (OCR) models
- Experiments in author handwriting identification and verification
- Tasks
- Image → text
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 4.2K
- Likes
- 37
Models trained on it
- leonnzhang/qwen3.5-0.8b-iam-25-GGUFImage text → text 156
- oguzhandurmaz/trocr-small-handwritten-onnx-int8Image → text 148
- eashwarsiddha/ocr-handwriting 20
- Kavic/nemotron-3.5-parse-handwriting-loraImage text → text 14
- eginugraha/paligemma-iam-line-qlora-adapterImage text → text 9
- savi8sant8s/monobyte-spelling-correctorText generation 0
- ismatsamadov/handwriting-recognition-iam 0
- ismatsamadov/handwriting_recognitionImage → text 0
How to load it
from datasets import load_dataset
ds = load_dataset("Teklia/IAM-line")