haskell
haskell copied to clipboard
Haskell bindings for TensorFlow
#111 started consolidating some packages; see also discussion in #66. We should consider doing this further, perhaps: - merge `tensorflow-logging` into `tensorflow` - merge `tensorflow-opgen` and/or `tensorflow-proto` into `tensorflow` There's...
Currently, to work around #92, we may do multiple unnecessary copies when fetching the result of a computation. (See also discussion in the PR #108.) If the output type requires...
There might be some docker-tuning that needs to be mentioned in `README.md`. ``` ... 179 more jobs ____[1,760 / 2,373] Compiling tensorflow/core/kernels/matching_files_op.cc ERROR: /tensorflow/tensorflow/core/kernels/BUILD:76:1: C++ compilation of rule '//tensorflow/core/kernels:strided_slice_op' failed:...
As an exercise to test my understanding, I've been trying to implement a simple dropout operation in the Haskell interface to Tensorflow. Following the example https://github.com/tensorflow/haskell/blob/master/tensorflow-mnist/app/Main.hs, I tried defining a...
If you do something like ```haskell TF.withNameScope "foo" $ return $ TF.add 1 2 ``` it will not set the name scope of the add op as expected. You have...
Many times errors from the underlying tensorflow implementation refer to the names of op nodes and it isn't obvious what haskell code generated those nodes. `withNameScope` can be used to...
Currently, the only options for fetching a multi-dimensional Tensor are lists or Vectors, both of which concatenate it into one dimension. (For example, a shape of [2,2,2] will turn into...
Support for Ampere architectures [was added in TensorFlow v2.4.0](https://github.com/tensorflow/tensorflow/releases/tag/v2.4.0). This architecture is used by Nividia's 3000 series. Currently, this project supports up to TensorFlow version v2.3.0. As a result, attempting...
I'm trying to load a TF2 SavedModel. I'm not sure if I'm doing it right, but I get the same error no matter what I do: ~~~ *** Exception: TensorFlowException...
Currently, Haskell is missing an easy-to-use ML library. I suspect Tensorflow might be Stack or Nix-bound for Haskell right now, but if someone were to update the bindings and provide...