Cydral
Cydral
> I've run the test on my GPU machine (`dtest --runall`) and they all pass. OK, so maybe it's due to the GPU I'm using (an A4500); could it be...
> Nice this is great. Fix up the minor stuff I commented on and I'll merge it right away :D Done.
> @Cydral This is all great. Really cool. Can I ask though, why do you use dlib instead of Pytorch for neural nets? Slowly adding transformer support to dlib is...
I'm continuing and making progress... I've just finished reworking gemm() to take into account the matricial dimension of 4D tensors.
There are no more conflicts for this modification. However, the precompilation tests fail because the tril_ class (previously committed in the master branch) is not found in the dnn.cpp program......
Something went wrong when I retrieved the latest versions after the integration of the tril_ layer... I just deleted the accesses to tril_ in the positional_encodings branch, in the dnn.ccp...
@davis, There were indeed new conflicts following the integration of the last class that you dealt with. It looks good now, but could you please go through the integration in...
Sorry for all these troubles, I've never had so many problems with merges on GitHub... I've just merged with the main branch and in my session I can see all...
It's OK now. The implementation in was in fact present but came from a previous merge (which explains why the tests and precompilation worked already)...