🌈 Welcome thread | 👥 Looking for teammates? | 🚀 Easy-2-Follow Code Notebooks

📝 Don't forget to participate in the Community Contribution Prize!

 Introduction

This is the most interesting puzzle in this Blitz! It utilizes your learnings from the previous puzzles. The challenge is simple, using a sound clip as an input can you accurately output only numbers mentioned in a text format? Feeling confused? Read on to know how to solve it.

Access an easy-to-use beginner-friendly code notebook over here.

💪 Getting Started

This puzzle will also use the DeepSpeech model for the purpose of sound prediction. The starter kit breaks down all the steps in a simple way. This notebook will guide you in installing, training, processing and predict the sound clips using Mozilla's DeepSpeech library.

Pretty simple, right? Check out the notebook and make your submission! 

💾 Dataset

In this dataset, you will need to predict what words were spoken in a voice. The sounds are in .wav the extension. The sample rate of the sound is 8000. 

📁 Files

Following files are available in the resources section:

• train.csv - (20000 samples) This CSV file contains a SoundID column as the file's name located in the corresponding folder name and a label column as the number spoked from the sound. 

• train.zip - (20000 samples) This CSV file containing sound files as the name corresponding to the SoundID column of the corresponding CSV file name.  

• val.csv - (2000 samples) This CSV file contains a SoundID column as the file's name located in the corresponding folder name and a label column as the number spoked from the sound. 

• val.zip - (2000 samples) This CSV file containing sound files as the name corresponding to the SoundID column of the corresponding CSV file name.

• test.csv - (5000 samples) This CSV file contains a SoundID column as the file's name located in the corresponding folder name and a label column as the number spoked from the sound. 

• test.zip - (5000 samples) This CSV file containing sound files as the name corresponding to the SoundID column of the corresponding CSV file name.

🚀  Submission

• Creating a submission directory
• Use test.csv and fill the corresponding labels. 
• Save the test.csv in the submission directory. The name of the above file should be submission.csv.
• Inside a submission directory, put the .ipynb notebook from which you trained the model and made inference and save it as original_notebook.ipynb.

         Overall, this is what your submission directory should look like - 

submission
├── submission.csv
└── original_notebook.ipynb

• Zip the submission directory!

Make your first submission here 🚀 !!

🖊 Evaluation Criteria

During the evaluation, the Mean Word Error Rate across each text in ground truth and submission will be used to test the efficiency of the model. We are using wer function from jiwer python library to calculate word error rate. 

🔗 Links

• 💪 Challenge Page: https://www.aicrowd.com/challenges/emotion-detection
• 🗣️ Discussion Forum: https://www.aicrowd.com/challenges/soundprediction/discussion
• 🏆 Leaderboard: https://www.aicrowd.com/challenges/soundprediction/leaderboards

📱 Contact

• Shubhamai