Comments (30)
Whya are you didn't response for my question 😔??
I have other things to do as well as support GitHub projects!
Some of your questions can be answered by reading and understanding the code in the notebook.
You download the speech commands dataset and uncompress it to the speech_data
folder. Follow the instructions at the top of the notebook:
Download data set
Download from: https://storage.cloud.google.com/download.tensorflow.org/data/speech_commands_v0.02.tar.gz - approx 2.3 GB
And then run
tar -xzf data_speech_commands_v0.02.tar.gz -C speech_data
This will give you a folder containing a list of words along with sample audio for each word.
The code runs over the list of words and extracts features for each word. If the word is the wake word - in our case "marvin" then we repeat it multiple times so that we have a balanced data set - approximately an equal number of wake words examples vs non-wakeword examples.
I've added an additional folder to the speech_data
folder with some background noise audio. Just create a folder in the speech_data
folder called _background_noise_
and put some recordings there of background noise.
If you don't have any background noise then comment out these lines:
for file_name in tqdm(get_files('_background_noise_'), desc="Processing Background Noise"):
process_background(file_name, words.index("_background"))
The same is true for _problem_noise_
. Create a folder called _problem_noise_
in the speech data folder and put audio that seems to confuse the neural network there.
Once again if you don't have any problem noise samples then comment out these lines:
for file_name in tqdm(get_files("_problem_noise_"), desc="Processing problem noise"):
process_problem_noise(file_name, words.index("_background"))
The same is the case for the _mar_sounds_
. I recorded a bunch of words that sound like "marvin" to help the neural network ignore similar words.
Comment out these lines if you don't have these samples:
for file_name in tqdm(get_files("_mar_sounds_"), desc="Processing problem noise"):
process_mar_sounds(file_name, words.index("_background"))
from diy-alexa.
I put the speech_data file to the model folder now it's working pretty thanks you solve my issue thank you
from diy-alexa.
ok i want to know are you use background audio files for training data??
from diy-alexa.
You want to train the neural network to detect your wake word. To do this it needs example audio files containing your wake word and example audio files that don't contain your wake word. These counter examples can be other words and background noise.
You also want it to be able to detect your wake word even when there is background noise. Or at least to be reasonably robust when there is background noise.
The background noise is used in the training of the neural network.
from diy-alexa.
You want to train the neural network to detect your wake word. To do this it needs example audio files containing your wake word and example audio files that don't contain your wake word. These counter examples can be other words and background noise.
You also want it to be able to detect your wake word even when there is background noise. Or at least to be reasonably robust when there is background noise.
The background noise is used in the training of the neural network.
ok how i training the neural network ?? using a jupyter notebook??
from diy-alexa.
You can use the jupyter notebooks in the repository, but if you want to use different training data then you will need to modify the code.
There are some good tutorials on the TensorFlow website that you can use to learn how it all works, but you are in for a lot work!
https://www.tensorflow.org/overview
from diy-alexa.
now I run the training data notebook but it'll occur an error "Processing problem noise" what is the reason??
from diy-alexa.
how do you add the audio sample to the training data??
from diy-alexa.
I recorded audio using my computer and built-in microphone to record additional background sounds.
You can comment out the lines that deal with the "mar" sounds if you don't need it.
from diy-alexa.
I recorded audio using my computer and built-in microphone to record additional background sounds.
You can comment out the lines that deal with the "mar" sounds if you don't need it.
I ask how you add audio sample to the Jupyter notebook
from diy-alexa.
I'm afraid the question doesn't make sense to me.
Are you asking how to read audio data into Python?
I think that you should follow some tutorials using jupyter notebook and tensorflow to learn more. The link I sent previously: https://www.tensorflow.org/overview is a good place to start.
There is also a good set of videos here: that will be very helpful for you to watch: https://www.youtube.com/playlist?list=PL-wATfeyAMNrtbkCNsLcpoAyBBRJZVlnf
from diy-alexa.
Ok i will watch the tutorial but i want to know are you add audio samples to the Jupyter notebook
from diy-alexa.
If you can't understand my question, let me know and I will try to explain my question a bit detail
from diy-alexa.
I don't understand your question.
from diy-alexa.
Are we need to add this data_speech_commands folder to the notebook then it will run through all these samples and output files?
from diy-alexa.
Whya are you didn't response for my question 😔??
from diy-alexa.
Are we need to add this data_speech_commands folder to the notebook then it will run through all these samples and output files?
I think you need to add command data file to the jupyter notebook ,friend . I also have this question .I also wait his reply
from diy-alexa.
I have other things to do as well as support GitHub projects!
Ok i know 🙂 i was thinked you still can't understand my question
from diy-alexa.
process_background(file_name
That's mean when i haven't background audio files then can i use this function
" for file_name in tqdm(get_files('background_noise'), desc="Processing Background Noise"):
process_background(file_name, words.index("_background")) " ??
And then run
tar -xzf data_speech_commands_v0.02.tar.gz -C speech_data
Where i run this ?? on command prompt ??
from diy-alexa.
If you don't have any background noise then comment out these lines:
for file_name in tqdm(get_files('_background_noise_'), desc="Processing Background Noise"): process_background(file_name, words.index("_background"))
When I have background files then how i add that background files to the code ??
from diy-alexa.
process_background(file_name
That's mean when i haven't background audio files then can i use this function " for file_name in tqdm(get_files('background_noise'), desc="Processing Background Noise"): process_background(file_name, words.index("_background")) " ??
And then run
tar -xzf data_speech_commands_v0.02.tar.gz -C speech_dataWhere i run this ?? on command prompt ??
You run this on the command prompt in the model folder. I don't think this will work in the normal windows command prompt. You would be better off installing WSL so that you have the full set of UNIX commands to work with.
from diy-alexa.
background_noise
Put the
If you don't have any background noise then comment out these lines:
for file_name in tqdm(get_files('_background_noise_'), desc="Processing Background Noise"): process_background(file_name, words.index("_background"))
When I have background files then how i add that background files to the code ??
Create a folder called background_noise
in the speech_data
folder and put the files there.
from diy-alexa.
Create a folder called
background_noise
in thespeech_data
folder and put the files there.
ok thanks
from diy-alexa.
Where i run this ?? on command prompt ??
You run this on the command prompt in the model folder. I don't think this will work in the normal windows command prompt. You would be better off installing WSL so that you have the full set of UNIX commands to work with.
Yes it's not work correctly in Windows command prompt,we can also use Ubuntu software on windows for this work
from diy-alexa.
Yes it's not work correctly in Windows command prompt,we can also use Ubuntu software on windows for this work
ok thanks friend
from diy-alexa.
I run this "Generate Training Data " notebook but it occurs an error "Processing Background Noise:0/0 [00:00<?it/s]0 0 0" what is the reason??
in your video, there is no occur that error
watch the difference between these screenshots what is the reason
from diy-alexa.
Why do you think it is happening? I don't know what is happening on your machine you need to debug the code and work out what the problem is.
It seems likely that it can't find the audio files. Have you downloaded the speech data and extracted it? Is it in the right place? Where is the code looking for the files?
What is the directory structure? You should have something like:
model/speech_data/
from diy-alexa.
Why do you think it is happening? I don't know what is happening on your machine you need to debug the code and work out what the problem is.
It seems likely that it can't find the audio files. Have you downloaded the speech data and extracted it? Is it in the right place?
I also think it's can't find the audio files.yes i downloaded the speech data and extract in right click on zip file and extract it.i not place anywhere that file. I just only download and extract it
from diy-alexa.
You need to create a folder called speech_data
in the model folder (where you are running the notebooks) and copy the extracted files into that folder.
It expects to have all the words from the commands dataset under the speech_data
folder.
from diy-alexa.
You need to create a folder called
speech_data
in the model folder (whIre you are running the notebooks) and copy the extracted files into that folder
How I copied the speech data folder to model folder?? Are you meaned notebook saved location to copy the speech data folder??
from diy-alexa.
Related Issues (20)
- how to implement the code to esp32 HOT 1
- Can you reuse pin for audio input and output ? HOT 3
- export tflite model to c++
- DIY Alexa not connecting to wifi HOT 1
- DIY Alexa not working
- DIY Alexa is not response HOT 6
- I CANT GET SOUND OUTPUT HOT 2
- Guru Meditation Error HOT 1
- Limitations to .WAV file? HOT 1
- INMP441 microphone issue
- Not an issue but more of a question on how to complie the code HOT 1
- detect wakeword process not working HOT 2
- cannot compile code HOT 1
- INMP441 not working HOT 5
- ADC-related compile errors when shifting to ESP32S3 HOT 5
- Tensorflow.io import gfile
- How to customize keyword on my dataset?
- How to custom Wake Word model ?
- Issue while running the Jupyter notebook Generate Training data HOT 1
- Program is resetting continuously HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from diy-alexa.