Expertise in Speech to Text / Natural Language Open Source Code
$25-50 CAD / hour
Kapalı
İlan edilme: 5 yıldan fazla önce
$25-50 CAD / hour
We are looking for an expert in Speech to Text / Natural Language Open Source Code. Our product's primary interface is with voice commands. We have a MEMS microphone and speaker.
Economically, it is simply not viable to use cloud services such as Google's, Alexa's, or IBM's Voice APIs.
The user interaction with our device is very simple and limited to only a few commands/actions. It should be possible to host our own OPEN SOURCE Speech to Text / Natural Language API and allow our own devices to connect to it to convert speech to text, determine the nature of the command, take appropriate action and then return a text that is converted to speech for the user to hear or text to initiate a pre-recorded message.
In addition, we would also like to provide the capability for customers to use Siri, Alexa, and/or Google Assistant through any of the devices they have at home, be it their iPhone, Android phone, Google Home or an Echo Dot, for example.
If you have experience and good knowledge on how to set this up, we would like to hear from you.
Thank you,
Sam
Have previously used almost all open source speech to text libraries. Would like to know more about the possible voice commands and desired level of accuracy etc. to check the feasibility basis the previous experience
Looking forward to hearing back from you !
Hi
We can make use of RNN for speech recognition.
Their are many different versions of LSTMs and GRUs for speech processing.
Lets discuss
I have been working on Machine Learning Regression, SVM, Decsion Trees and Deep Learning RNN, CNN and GANs
i have worked in Natural language processing in text and speech both using different API like Freetts and Sphinx. also worked on different text processing algorithms.
I have a few queries:
1. Voice recognition can depend on accents it is difficult to recognize all the accents. So do you have enough data points to train the system for the target accent?
2. You mentioned about MEMS microphone and speaker but you did not tell about the Computational unit - Is it going to be a full blown computer or a miniature computer unit such as Raspberry PI?
**** This is an important factor to decide on the open source speech recognition system ****
3. You mentioned about the capability to use Siri etc. For that we just need to tap into their APIs.
I am Pranav Patil, representing Iristechsys Software Services LLP, an ISO 9001:2015 company based out of Pune, India.
My Linked-In profile will provide you overview of my experience.
My experience is mostly in Convutional Neural Networks for Image processing but the foundation for speech synthesis remains the same, if you plan to use DNNs.
Prima facie I think SPHINX from CMU would be the most suitable for your approach. Alternatively if you can afford a heavy weight processing engine in your product then it is better to go with Deep Neural Networks using Kaldi.
Hi Dear, We are currently working on Alexa voice search application in Android so I can easily develop your application, I'm ready to start your project.
We don't just build an application, We build the reputation, trust, and relationship.
I have excellent experience in Application development. We have 5+ years experience in application development and 7+ years in Information technology field.
I have rich experiences in android and ios area. I have developed several design, content management, and social apps.I am sure I can do your project within your timeline and budget.
We have 5 Star rating in freelancer in application development and, We have 90+ Android and iOS application developed.
I have gone through your job description very carefully and We are very much interested in your project and we are enough confident to start work on your project.
I have a few queries and suggestions to discuss. Please initiate the Freelancer Chat Box
Looking forward to hearing from you.
Thanks
Kumarsingh
I am working on speech products, weather it is Speech to Text or Text to Speech, from last few years. I have experience of doing it by using open source tools like Kaldi, Tensorflow, Pytorch and Python. I have successfully deployed the solution on client ends.
So I think I can build on premise speech solution without going to any third party APIs like watson etc.