Hi HN, I am Anubhav from RamanLabs. We have been developing end to end Computer-vision modules to make it easy for developers, hobbyists to integrate such functionality into their applications with minimal amount of code.
All modules are developed to run real-time on consumer-grade CPUs[0]. For now we are releasing only Python-language SDK. Demos are provided to allow users to test performance on desired data-distribution.
Framework powering these modules in completely written in Nim language, which under the hood wraps some Operation's implementation provided by libraries like OpenBLAS/CuBlas.
We also maintain a blog[1] where we hope to share more underlying technical details.
Hi Anubhav, are all modules built by your team? Since if it were based on already available models/repos, then I can't find licenses/attr for the same.
Almost All models are based on the Darknet[0] configurations which would come under this License[1]. We further trained/finetuned those architectures on additional datasets collected by us.
Whole framework actual used for final products is completely written in Nim.
Comments
Hi HN, I am Anubhav from RamanLabs. We have been developing end to end Computer-vision modules to make it easy for developers, hobbyists to integrate such functionality into their applications with minimal amount of code.
All modules are developed to run real-time on consumer-grade CPUs[0]. For now we are releasing only Python-language SDK. Demos are provided to allow users to test performance on desired data-distribution.
Framework powering these modules in completely written in Nim language, which under the hood wraps some Operation's implementation provided by libraries like OpenBLAS/CuBlas.
We also maintain a blog[1] where we hope to share more underlying technical details.
[0] Quad-core CPU with AVX2 instructions.
[1] https://ramanlabs.in/static/blog
[2] https://ramanlabs.in/static/tutorial
Hi Anubhav, are all modules built by your team? Since if it were based on already available models/repos, then I can't find licenses/attr for the same.
Hi,
Almost All models are based on the Darknet[0] configurations which would come under this License[1]. We further trained/finetuned those architectures on additional datasets collected by us.
Whole framework actual used for final products is completely written in Nim.
[0] https://github.com/AlexeyAB/darknet [1] https://github.com/AlexeyAB/darknet/blob/master/LICENSE