Bhashini Model Enhancements

[rohitsamagra]

• Accuracy to checked using Fuzzy Search + Bhashini Approach -- Gautam is checking the script for how it is working on the children, difference between actual Bhashini and Steno result. Do we need the training or not?
• If not accurate after fuzzy search, train the model - do we have the data in the format required? -- Gautam's sheet -- Aiming for 90-95%

[rohitsamagra]

• 89% with only Bhashini Model
• 95% with fuzzy
• Training might not be required
• Preet/Gautam to share in a well required format

[charanpreet-s]

@charanpreet-s to send mail with final results