Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / coreylowman/dfdx issues and pull requests
#928 - Safetensor panics instead of returning Err
Issue -
State: open - Opened by opfromthestart about 1 month ago
#927 - Make tests pass on stable + support Pool2D & Conv{1,2}d
Pull Request -
State: open - Opened by twitchyliquid64 2 months ago
#926 - Tests failing on HEAD /w cargo 1.78.0
Issue -
State: open - Opened by twitchyliquid64 3 months ago
#925 - Kernels written in rust-gpu
Issue -
State: open - Opened by LegNeato 6 months ago
- 1 comment
#924 - can not perform tensor operations (add/sub/mul/...) on tensors with non float datatype
Issue -
State: open - Opened by lucomotive 6 months ago
#923 - fix: small typo in advanced-train-loop.rs
Pull Request -
State: open - Opened by LeoDog896 7 months ago
#922 - docs(readme): fix examples hyperlink
Pull Request -
State: open - Opened by mrtnvgr 7 months ago
#921 - Non-trainable parameters?
Issue -
State: open - Opened by hovinen 7 months ago
- 2 comments
#920 - Avoid CI errors
Pull Request -
State: open - Opened by swfsql 7 months ago
#919 - Added `TryUnstack` for tensors.
Pull Request -
State: open - Opened by swfsql 8 months ago
#918 - Add Mamba (minimal)
Pull Request -
State: closed - Opened by swfsql 8 months ago
- 1 comment
#917 - Add split_tensor_along method
Pull Request -
State: open - Opened by swfsql 8 months ago
#916 - Add RMS normalization
Pull Request -
State: open - Opened by swfsql 8 months ago
#915 - Add SiLU activation function
Pull Request -
State: open - Opened by swfsql 8 months ago
#914 - Impl `core::ops::Sub` for Dim types
Pull Request -
State: open - Opened by swfsql 8 months ago
#913 - Safetensors changes
Pull Request -
State: open - Opened by swfsql 8 months ago
#912 - Remove functions calling the deprecated `_MM_SET_FLUSH_ZERO_MODE`
Pull Request -
State: open - Opened by rainiwu 8 months ago
- 3 comments
#911 - make non-const conv1d available to stable rust
Pull Request -
State: closed - Opened by optman 8 months ago
#910 - CUDA kernels missing __hmin and __hmax
Issue -
State: open - Opened by opfromthestart 9 months ago
- 1 comment
#909 - Add missing backtick (`) in crate-level documentation
Pull Request -
State: closed - Opened by rainiwu 9 months ago
#908 - OpenXLA Support
Issue -
State: open - Opened by vyeevani 9 months ago
- 3 comments
#907 - Unable to build with old CUDA version (`CUDA_COMPUTE_CAP = 52`)
Issue -
State: open - Opened by nikvoid 9 months ago
#906 - feat(wgpu): add to_dtype kernel
Pull Request -
State: closed - Opened by DonIsaac 9 months ago
#905 - How does one update one model from another model?
Issue -
State: closed - Opened by hovinen 9 months ago
- 1 comment
#904 - feat(webgpu): add tensor operations
Pull Request -
State: open - Opened by DonIsaac 9 months ago
- 1 comment
#903 - Add `nn::AdaptiveAvgPool2D`
Issue -
State: open - Opened by quietlychris 9 months ago
#902 - Gradient Clipping
Pull Request -
State: open - Opened by swfsql 10 months ago
#901 - small doc typo fix
Pull Request -
State: closed - Opened by HyperCodec 10 months ago
#900 - Unclear how to handle error type in `dfdx::nn::LoadFromNpz::load`
Issue -
State: open - Opened by hovinen 10 months ago
#899 - Run tests with miri
Pull Request -
State: closed - Opened by swfsql 10 months ago
- 1 comment
#898 - Reduce test sizes
Issue -
State: closed - Opened by swfsql 10 months ago
- 1 comment
#897 - Run tests with miri
Issue -
State: closed - Opened by swfsql 10 months ago
- 1 comment
#896 - Implement AbsKernelOp for WebGPU backend
Pull Request -
State: closed - Opened by favilo 10 months ago
- 7 comments
#895 - Add `Prodigy` optimizer
Pull Request -
State: closed - Opened by swfsql 10 months ago
- 1 comment
#894 - Add `Prodigy` optimizer
Issue -
State: closed - Opened by swfsql 10 months ago
- 1 comment
#893 - Implement Webgpu device
Pull Request -
State: closed - Opened by favilo 10 months ago
- 4 comments
#892 - Split `TryConcatAlong` into different traits
Pull Request -
State: closed - Opened by swfsql 11 months ago
- 3 comments
#891 - Split `TryConcatAlong` into different traits
Issue -
State: closed - Opened by swfsql 11 months ago
#890 - Adds `OUTPUT_PADDING` to `ConvTrans2D`
Pull Request -
State: open - Opened by swfsql 11 months ago
#889 - Add `OUTPUT_PADDING` to `ConvTrans2D`
Issue -
State: open - Opened by swfsql 11 months ago
#888 - Send/Sync for Device
Issue -
State: open - Opened by blogle 11 months ago
- 1 comment
#887 - On older GPU there are no hcos/hsin/hlog/hexp
Pull Request -
State: closed - Opened by optman 11 months ago
- 1 comment
#886 - Replace explicit features and paths on generated code
Pull Request -
State: closed - Opened by swfsql 11 months ago
#885 - Replace explicit features and paths on generated code
Issue -
State: closed - Opened by swfsql 11 months ago
#884 - Add `forward_mut` in `Sequential` derive
Pull Request -
State: closed - Opened by swfsql 11 months ago
#883 - Input Wrapper
Pull Request -
State: closed - Opened by swfsql 11 months ago
- 2 comments
#882 - make device capable of stacking and concatenating usize type
Pull Request -
State: closed - Opened by optman 11 months ago
#881 - Bug: `Sequential` macro provide `forward_mut` as `forward`
Issue -
State: closed - Opened by uselessgoddess 11 months ago
#880 - Examples or resources for autodiff with 2 networks?
Issue -
State: open - Opened by Cobrand 11 months ago
#879 - Question / clarification regarding heap allocations
Issue -
State: closed - Opened by emchristiansen 11 months ago
- 2 comments
#878 - Consider helpers for accessing tensors from tuples and input wrappers
Issue -
State: closed - Opened by swfsql 11 months ago
- 1 comment
#877 - Bump dependency versions
Pull Request -
State: closed - Opened by coreylowman 11 months ago
#876 - Allocate 1 byte for workspace instead of 0
Pull Request -
State: closed - Opened by coreylowman 11 months ago
#875 - [Breaking] Combining separate device errors into single `dfdx::tensor::Error` enum
Pull Request -
State: closed - Opened by coreylowman 11 months ago
#874 - trait TryConcatAlong not satisfied when using constants
Issue -
State: closed - Opened by opfromthestart 12 months ago
- 1 comment
#873 - Implementations for u8->u64 and i8->i64
Pull Request -
State: closed - Opened by timwedde 12 months ago
- 5 comments
#872 - Unnecessary loss of precision when computing loss functions
Issue -
State: closed - Opened by ariasanovsky 12 months ago
- 2 comments
#871 - Different results when CPU feature is on vs off
Issue -
State: closed - Opened by jafioti 12 months ago
- 2 comments
#870 - Alloc zero size memory on old model GPU may fail.
Issue -
State: closed - Opened by optman almost 1 year ago
#869 - Sigmoid with residue
Pull Request -
State: closed - Opened by opfromthestart about 1 year ago
- 3 comments
#868 - Making ToDtypeKernel pub
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#867 - ConcatAlong is very slow
Issue -
State: closed - Opened by opfromthestart about 1 year ago
- 12 comments
Labels: gpu, optimization
#866 - Device doesn't cover all traits from Cpu/Cuda
Issue -
State: open - Opened by jackerschott about 1 year ago
- 2 comments
#865 - Is there an equivalent to torch.masked_select?
Issue -
State: closed - Opened by jackerschott about 1 year ago
- 1 comment
#864 - [Breaking] Changing rhs for binary ops with scalars to `Into<f64>` instead of `E`
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#863 - Adds 1d convolutions
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
- 1 comment
#862 - Convenience of mixed tensor scalar operations
Issue -
State: closed - Opened by jackerschott about 1 year ago
- 1 comment
#861 - Question: How to load checkpoints?
Issue -
State: closed - Opened by rjzak about 1 year ago
- 1 comment
#860 - feat: publish stacked API
Pull Request -
State: closed - Opened by LeoDog896 about 1 year ago
#859 - Expose AddDim to public API
Issue -
State: closed - Opened by LeoDog896 about 1 year ago
- 1 comment
#858 - feat: N-sized tuple (2 <= n <= 8)
Pull Request -
State: open - Opened by LeoDog896 about 1 year ago
#857 - Grouped Linear layers
Issue -
State: open - Opened by opfromthestart about 1 year ago
- 1 comment
#856 - Deriving Shape for C like enums
Issue -
State: closed - Opened by ZacharyKamerling about 1 year ago
- 1 comment
#855 - fixed __CUDA_ARCH__ check for compat with cuda 12.2
Pull Request -
State: closed - Opened by clstatham about 1 year ago
- 1 comment
#854 - [Breaking] Rewrite of `nn` to enable runtime layer sizes, proc macro declarations, and more
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
- 1 comment
#853 - Fixing nightly should_panic messages on tests
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#852 - [Breaking] Removing `SafeDtype`, adding byte conversion traits to Unit to make safetensors serialization easier
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#851 - [Breaking] Removing extra safetensors Error struct
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#850 - How to calculate loss to maximize a value
Issue -
State: closed - Opened by opfromthestart about 1 year ago
- 3 comments
#849 - Add complex numbers
Pull Request -
State: open - Opened by opfromthestart about 1 year ago
- 3 comments
#848 - Differing behavior between Pytorch and dfdx
Issue -
State: closed - Opened by opfromthestart about 1 year ago
- 6 comments
#847 - Adds Gradients::retain_current_grads_as_leafs
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#846 - Fixing bug with 0 sized tensors
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#845 - Expose TensorLike and Gradients::get_ref_checked
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#844 - Bug: `as_vec` crashes for empty Tensors
Issue -
State: closed - Opened by emchristiansen about 1 year ago
#843 - `TapeGlobal` and thoughts on the tape variants
Issue -
State: open - Opened by emchristiansen about 1 year ago
- 3 comments
#842 - Bug: Incompatible uniqueness assumption causes crashes
Issue -
State: open - Opened by emchristiansen about 1 year ago
- 1 comment
#841 - The merge semantics for `Arc<Mutex<OwnedTape<_, _>>>` seem a bit unintuitive
Issue -
State: open - Opened by emchristiansen about 1 year ago
- 2 comments
#840 - Please minimize the requirements for the optimizers
Issue -
State: closed - Opened by emchristiansen about 1 year ago
- 1 comment
#839 - Impossible to create a module with a parameter that lacks a const shape?
Issue -
State: closed - Opened by emchristiansen about 1 year ago
- 1 comment
#838 - [Breaking] Adding `type Output` to TryAdd/TryDiv/TryMul/TrySub traits
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#837 - Consider adding a `Tape` type where the tape is optional, e.g. `Option<Arc<Mutex<OwnedTape<_, _>>>>`
Issue -
State: closed - Opened by emchristiansen about 1 year ago
- 2 comments
#836 - impl `Merge` for `NoneTape` for `OwnedTape` and `Arc<Mutex<OwnedTape>>`
Issue -
State: closed - Opened by emchristiansen about 1 year ago
- 1 comment
#835 - impl Tape for Arc<Mutex<OwnedTape>>
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
- 1 comment
#834 - Adding documentation to dtypes module and amp
Pull Request -
State: closed - Opened by coreylowman about 1 year ago
#833 - `ArcTape`: Add another `Tape` impl?
Issue -
State: closed - Opened by emchristiansen about 1 year ago
- 4 comments
#832 - Add Gradient scaler
Issue -
State: open - Opened by coreylowman about 1 year ago
- 1 comment
#831 - Adds `unstack`
Pull Request -
State: open - Opened by emchristiansen about 1 year ago
- 1 comment
#830 - Accurate gelu cuda fix
Pull Request -
State: closed - Opened by jcrist1 about 1 year ago
- 2 comments
#829 - Fixes compilation issues with accurate gelu
Pull Request -
State: closed - Opened by coreylowman about 1 year ago