jsalbert / sound_classification_ml_production

In this repository you will find an end to end hands-on tutorial of an example of machine learning in production. The objective will be to create and deploy in the cloud a machine learning application able to recognize and classify different audio sounds.

Hi there, I'm Albert 👋

I am a machine learning engineer with extensive experience in deep learning, computer vision, natural language processing, generative models, and audio processing.

📜 My Values

🌟 Open minded and curious

⚡️ Passionate about solving complex problems

🍏 Honest and transparent

👨🏻‍💻 Lifelong learner

✍🏻 Latest Blog Posts

• Feedback Alignment Methods
• Launch Your ML App in 4 Easy Steps
• How to Parallelize Your Python Tests Easily with CircleCI

📬 Get in touch

           

sound_classification_ml_production's Issues

Query

Hi,
I have a doubt as of now i don't have cloud access what should i do dude

Is my audio file correct, like the others in folder 0-10?

Hi, I add some audio files and named it as folder 11, and ran code UrbanSound8k_data_exploration.ipynb. I found that my audio is from 0-1024 Hz, while the other UrbanSound8k audio is from 0-8192 Hz.

So is my audio correct enough for training? What's the difference? thank you.

my audio's mfcc graph (mine is on right side):

one of my audio:
dashengdahan-41c342fc746bac53b7ddf552bed57e4214d67092.[0.93-2.59].wav.zip

How to produce .npy features file?

Hi, I run you code well, But found that .npy is produced before and just download. So I want to know how to produce my .npy files? I run this "fe.compute_save_features", but only got .png feature files. How can I produce .npy like you did, thx!