Update README.md

README.md

@@ -1,199 +1,83 @@
 ---
 library_name: transformers
-tags: []
+license: cc-by-4.0
+language:
+  - smi
+  - smj
+  - sme
+  - sma
+  - smn
+  - nor
+datasets:
+  - Sprakbanken/synthetic_sami_ocr_data
+base_model:
+  - microsoft/trocr-base-printed
 ---
 
-# Model Card for Model ID
+# Model Card for Sprakbanken/trocr_smi_nor_pred_synth
+This is a TrOCR-model for OCR (optical character recognition) of Sámi languages.
+It can be used to recognize text in images of printed text (scanned books, magazines, etc.) in North Sámi, South Sámi, Lule Sámi, and Inari Sámi.
 
-<!-- Provide a quick summary of what the model is/does. -->
+## Collection details 
+This model is a part of our collection of OCR models for Sámi languages.
 
+The following TrOCR models are available:
+- [Sprakbanken/trocr_smi](https://huggingface.co/Sprakbanken/trocr_smi): [microsoft/trocr-base-printed](https://huggingface.co/microsoft/trocr-base-printed) fine-tuned on manually annotated Sámi data
+- [Sprakbanken/trocr_smi_nor](https://huggingface.co/Sprakbanken/trocr_smi_nor): microsoft/trocr-base-printed fine-tuned on manually annotated Sámi and Norwegian data
+- [Sprakbanken/trocr_smi_pred](https://huggingface.co/Sprakbanken/trocr_smi_pred): microsoft/trocr-base-printed fine-tuned on manually annotated and automatically transcribed Sámi data
+- [Sprakbanken/trocr_smi_nor_pred](https://huggingface.co/Sprakbanken/trocr_smi_nor_pred): microsoft/trocr-base-printed fine-tuned on manually annotated and automatically transcribed Sámi data, and manually annotated Norwegian data
+- [Sprakbanken/trocr_smi_synth](https://huggingface.co/Sprakbanken/trocr_smi_synth): microsoft/trocr-base-printed fine-tuned on [Sprakbanken/synthetic_sami_ocr_data](https://huggingface.co/datasets/Sprakbanken/synthetic_sami_ocr_data), and then on manually annotated Sámi data
+- [Sprakbanken/trocr_smi_pred_synth](https://huggingface.co/Sprakbanken/trocr_smi_pred_synth): microsoft/trocr-base-printed fine-tuned on Sprakbanken/synthetic_sami_ocr_data, and then fine-tuned on manually annotated and automatically transcribed Sámi data
+- [Sprakbanken/trocr_smi_nor_pred_synth](https://huggingface.co/Sprakbanken/trocr_smi_nor_pred_synth): microsoft/trocr-base-printed fine-tuned on Sprakbanken/synthetic_sami_ocr_data, and then fine-tuned on manually annotated and automatically transcribed Sámi data, and manually annotated Norwegian
 
+[Sprakbanken/trocr_smi_pred_synth](https://huggingface.co/Sprakbanken/trocr_smi_pred_synth) is the model that achieved the best results (of the TrOCR models) on our test dataset.
 
 ## Model Details
+This model is TrOCR-printed base model trained on Sprakbanken/synthetic_sami_ocr_data for 5 epochs,
+and then fine-tuned on manually annotated and automatically transcribed Sámi data, and manually annotated Norwegian. See our paper for more details.
 
-### Model Description
-
-<!-- Provide a longer summary of what this model is. -->
-
-This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
 
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
+### Model Description
 
-### Model Sources [optional]
+- **Developed by:** The National Library of Norway
+- **Model type:** TrOCR
+- **Languages:**  North Sámi (sme), South Sámi (sma), Lule Sámi (smj), and Inari Sámi (smn)
+- **License:** [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)
+- **Finetuned from model :** [TrOCR-printed base model](https://huggingface.co/microsoft/trocr-base-printed)
 
-<!-- Provide the basic links for the model. -->
+### Model Sources 
 
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
+- **Repository:** https://github.com/Sprakbanken/nodalida25_sami_ocr
+- **Paper:** "Enstad T, Trosterud T, Røsok MI, Beyer Y, Roald M. Comparative analysis of optical character recognition methods for Sámi texts from the National Library of Norway. Accepted for publication in Proceedings of the 25th Nordic Conference on Computational Linguistics (NoDaLiDa) 2025." (preprint coming soon.)
 
 ## Uses
-
-<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
-
-### Direct Use
-
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
-
-[More Information Needed]
-
-### Downstream Use [optional]
-
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
-
-[More Information Needed]
+You can use the raw model for optical character recognition (OCR) on single text-line images in North Sámi, South Sámi, Lule Sámi, and Inari Sámi. 
 
 ### Out-of-Scope Use
+The model only works with images of lines of text.
+If you have images of entire pages of text, you must segment the text into lines first to benefit from this model.
 
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-
-[More Information Needed]
-
-## Bias, Risks, and Limitations
-
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-
-[More Information Needed]
-
-### Recommendations
-
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
 
 ## How to Get Started with the Model
 
 Use the code below to get started with the model.
 
-[More Information Needed]
-
-## Training Details
-
-### Training Data
-
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
-
-[More Information Needed]
-
-### Training Procedure
-
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-
-#### Preprocessing [optional]
-
-[More Information Needed]
-
-
-#### Training Hyperparameters
-
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
-
-#### Speeds, Sizes, Times [optional]
-
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
-
-[More Information Needed]
-
-## Evaluation
-
-<!-- This section describes the evaluation protocols and provides the results. -->
-
-### Testing Data, Factors & Metrics
-
-#### Testing Data
-
-<!-- This should link to a Dataset Card if possible. -->
+```python
+from transformers import TrOCRProcessor, VisionEncoderDecoderModel
+from PIL import Image
 
-[More Information Needed]
+processor = TrOCRProcessor.from_pretrained("Sprakbanken/trocr_smi_pred_synth")
+model = VisionEncoderDecoderModel.from_pretrained("Sprakbanken/trocr_smi_pred_synth")
 
-#### Factors
+image = Image.open("path_to_image.jpg").convert("RGB")
 
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+pixel_values = processor(image, return_tensors="pt").pixel_values
+generated_ids = model.generate(pixel_values)
 
-[More Information Needed]
-
-#### Metrics
-
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
-
-[More Information Needed]
-
-### Results
-
-[More Information Needed]
-
-#### Summary
-
-
-
-## Model Examination [optional]
-
-<!-- Relevant interpretability work for the model goes here -->
-
-[More Information Needed]
-
-## Environmental Impact
-
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
-
-Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
-
-- **Hardware Type:** [More Information Needed]
-- **Hours used:** [More Information Needed]
-- **Cloud Provider:** [More Information Needed]
-- **Compute Region:** [More Information Needed]
-- **Carbon Emitted:** [More Information Needed]
-
-## Technical Specifications [optional]
-
-### Model Architecture and Objective
-
-[More Information Needed]
-
-### Compute Infrastructure
-
-[More Information Needed]
-
-#### Hardware
-
-[More Information Needed]
-
-#### Software
-
-[More Information Needed]
-
-## Citation [optional]
-
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
-
-**BibTeX:**
-
-[More Information Needed]
+generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+```
+## Citation 
 
 **APA:**
 
-[More Information Needed]
-
-## Glossary [optional]
-
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
-
-[More Information Needed]
-
-## More Information [optional]
-
-[More Information Needed]
-
-## Model Card Authors [optional]
-
-[More Information Needed]
-
-## Model Card Contact
-
-[More Information Needed]
+Enstad, T., Trosterud, T., Røsok, M. I., Beyer, Y., & Roald, M. (2025). Comparative analysis of optical character recognition methods for Sámi texts from the National Library of Norway. Proceedings of the 25th Nordic Conference on Computational Linguistics (NoDaLiDa).