mlx
mlx copied to clipboard
ml-explore
MLX: An array framework for Apple silicon
Not sure if this is an issue or I'm just holding it wrong :-) I've got an alternative approach that seems to work with gradients by building up my output...
It would also be nice to have a binary for this on conda-forge. You can see my start at implementing a recipe for this at https://github.com/conda-forge/staged-recipes/pull/24687/files As the CI over...
I'm working on the DQN code but overall it's still 10x slower than torch. I've narrowed the issue down to a line with only an `argmax`. Can anyone help me...
System details : Sonoma 14.2.1 xcode-select version 2405 I am trying to run the default example provided by mlx-examples directory. `python mistral.py --model_path Mistral-7B-Instruct-v0.2/ --prompt "My name is"` I am...
Block-diagonal attention allows keys/queries from within the same sequence to attend to each other and not other sequences, even when multiple sequences are packed/concatenated together. This is really useful to...
PyTorch Navie Implementation: https://github.com/pytorch/pytorch/blob/main/aten/src/ATen/native/NaiveConvolutionTranspose2d.cpp PyTorch Navie Cuda Implementation: https://github.com/pytorch/pytorch/blob/main/aten/src/ATen/native/cuda/NaiveConvolutionTranspose2d.cu I guess so this is the metal implementation of Conv Tranpose in PyTorch: https://github.com/pytorch/pytorch/blob/015bd0e0a189f929e469c6bc75fe1541c18a014d/aten/src/ATen/native/mps/operations/Convolution.mm#L709
When building the mlx project I'm getting an error in regards to : NEON intrinsics not available with the soft-float ABI Please use -mfloat-abi=softfp or -mfloat-abi=hard ChatGPT suggests updating MakeFile...
It looks like this is still missing many matrix operations like QR, SVD, einsum, etc. Is there a clear path to using these with or without MLX? [This has been...
Grouped convolutions are awesome! Would love to be able to use them in MLX. :)
The docs have more information on [troubleshooting installation](https://ml-explore.github.io/mlx/build/html/install.html). Two of the more common issues are described below (as well as in the docs). # Installing from PyPI A common problem...
