Information extraction from Cheshire fire incident reports using Mistral

[SurajBaloni]

This PR adds sample notebook on - Information extraction from Cheshire fire incident reports using Mistral language model.

#2115

Table Of Contents location:
Deep Learning > NLP > Information extraction from Cheshire fire incident reports using Mistral

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[SurajBaloni]

@BP-Ent Please review this sample notebook

[jyaistMap]

• looks good visually from my review, but cannot speak to technical accuracy
• local build of doc appears like this for half of doc, rest of it looks good
• the right-hand navigation all works

[jyaistMap]

@priyankatuteja

• when you approve this, I can merge python-api-doc #288 to update the Table of Contents to reflect new sample

[BP-Ent]

• Refer to the section Install deep learning dependencies of the arcgis.learn module for detailed documentation on installing the dependencies.

• To learn more about how EntityRecognizer works, please refer to the guide on Named Entity Extraction with arcgis.learn.

Reply via ReviewNB

[BP-Ent]

Data preparation involves splitting the data into training and validation sets, creating the necessary data structures for loading data into the model. The prepare_data() function can directly read the training samples in one of the above specified formats and automate the entire process.

Reply via ReviewNB

[BP-Ent]

The EntityRecognizer model in arcgis.learn can be used with Hugging Face Transformers or with large language model backbones. For this sample use case, we will use the Mistral model backbone to extract entities from the text.

Run the command below to see what backbones are supported for the entity recognition task.

Reply via ReviewNB

[BP-Ent]

First we will create the model using the EntityRecognizer() constructor and passing the following parameters:

data: The databunch created using the prepare_textdata method.

backbone: To use mistral as the model backbone, use backbone="mistral".

prompt: Text string describing the task and its guardrails. This is an optional parameter.

Reply via ReviewNB

[BP-Ent]

Important metrics to look at while measuring the performance of the EntityRecognizer model are Precision, Recall, and F1-measures.

Reply via ReviewNB

[BP-Ent]

To find precision, recall, and f1 scores per label/class, we will call the model's metrics_per_label() method.

Reply via ReviewNB

[BP-Ent]

Now that we have the trained model, let's look at how the model performs.

Reply via ReviewNB

[BP-Ent]

The load() method takes the path to the .emd file as a required argument.

Reply via ReviewNB

[BP-Ent]

Now we can use the trained model to extract entities from new text documents using the extract_entities() method. This method expects either the folder path of where the new text documents are located or a list of text documents.

Reply via ReviewNB

[BP-Ent]

Suggestions made on reviewnb

done