Update README.md

README.md

@@ -34,12 +34,12 @@ model-index:
       value: 0.9916725247390905
 ---
 
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
 
 # nutrition-extractor
 
 This model is a fine-tuned version of [microsoft/layoutlmv3-large](https://huggingface.co/microsoft/layoutlmv3-large) on the openfoodfacts/nutrient-detection-layout dataset.
+It allows to automatically extract nutrition values from images of nutrition tables.
+
 It achieves the following results on the evaluation set:
 - Loss: 0.0534
 - Precision: 0.9545
@@ -49,15 +49,25 @@ It achieves the following results on the evaluation set:
 
 ## Model description
 
-More information needed
+This model can extract nutrient values from nutrition tables. This was developped as part of the Nutrisight project.
+
+For more information about the project, please refer to the [nutrisight](https://github.com/openfoodfacts/openfoodfacts-ai/tree/develop/nutrisight) directory in the openfoodfacts-ai GitHub repository.
+
+As any model using the LayoutLM architecture, this model expects as input:
+
+- the image
+- the tokens (string) on the images
+- the 2D coordinates of each tokens
+
+The tokens and their 2D position is provided by an OCR model. This model was trained using OCR results coming from Google Cloud Vision.
 
 ## Intended uses & limitations
 
-More information needed
+This model is only intended to be used on images of products where a nutrition table can be found.
 
 ## Training and evaluation data
 
-More information needed
+The training and evaluation data can be found on the [dataset page](https://huggingface.co/datasets/openfoodfacts/nutrient-detection-layout).
 
 ## Training procedure