jatinchowdhury18/rtneural issues and pull requests

#159 - Testing compatibility with half-precision floating point types

Pull Request - State: closed - Opened by jatinchowdhury18 2 months ago - 1 comment

#158 - Add getLayers method to ModelT

Pull Request - State: closed - Opened by jatinchowdhury18 2 months ago

#157 - Large GRU Segmentation Fault

Issue - State: open - Opened by MattyAB 3 months ago - 3 comments

#156 - torch_helpers: Add layer_index argument to loadGRU and loadLSTM

Pull Request - State: closed - Opened by jatinchowdhury18 3 months ago

#155 - Dense layers with optionally no bias

Pull Request - State: closed - Opened by jatinchowdhury18 3 months ago

#154 - Dense layers without bias

Issue - State: closed - Opened by mikeoliphant 3 months ago - 1 comment

#153 - Add nam-lv2 and stompbox to "powered by" section of README

Pull Request - State: closed - Opened by mikeoliphant 3 months ago - 1 comment

#152 - Add MathsProvider to template for dynamic model loading

Pull Request - State: closed - Opened by mikeoliphant 3 months ago - 3 comments

#151 - Only include CPM for testing

Pull Request - State: closed - Opened by jatinchowdhury18 3 months ago

#150 - Ensure MathsProvider template arguments are forwarded through ModelT::parseJson

Pull Request - State: closed - Opened by jatinchowdhury18 3 months ago

#149 - Use tanh to speed up sigmoid calculation

Pull Request - State: closed - Opened by mikeoliphant 3 months ago - 10 comments

#148 - Support case where "num_layers" of 'subnetwork' > 1 in torch_helpers

Issue - State: closed - Opened by nvssynthesis 4 months ago - 2 comments

#147 - No 'vanilla' RNN layer support?

Issue - State: open - Opened by nvssynthesis 4 months ago - 2 comments

#144 - Strides argument in RTNeural's Conv1D layers

Issue - State: closed - Opened by ABSounds 7 months ago - 6 comments

#143 - Add support for NAM files

Issue - State: closed - Opened by christoph-hart 7 months ago - 39 comments

#139 - RTNEURAL_DEFAULT_ALIGNMENT=8 on armv7, EIGEN backend

Issue - State: closed - Opened by MaxPayne86 9 months ago - 6 comments

#136 - bfloat16 Support

Issue - State: closed - Opened by tarlanahad 9 months ago - 4 comments

#132 - Benchmark results

Issue - State: closed - Opened by 7sharp9 11 months ago - 1 comment

#131 - Offline build

Issue - State: closed - Opened by hfiguiere 11 months ago

#100 - AIDA-X now does AU as well

Pull Request - State: closed - Opened by falkTX almost 2 years ago - 1 comment

#99 - Is it possible to use a batch of data as input?

Issue - State: closed - Opened by CesareTonola almost 2 years ago - 1 comment

#98 - Add AIDA-X to "Powered by RTNeural" project list

Pull Request - State: closed - Opened by falkTX almost 2 years ago

#97 - Allow Torch layers to be loaded without bias

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago

#96 - Add helpers for loading layer weights exported from PyTorch

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago - 1 comment

#95 - Add example for loading a GRU layer from PyTorch

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago - 1 comment

#94 - Update XSIMD submodule

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago

#93 - Add compile option to exclude <iostream> header

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago - 1 comment

#92 - Add Conv2D and BatchNorm2D implementations for XSIMD

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago

#91 - STL implementations for Conv2D and BatchNorm2D

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago - 1 comment

#90 - Add example for Conv1D exported from PyTorch

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago - 1 comment

#89 - Streaming Conv 2D and Batchnorm 2D implementations

Pull Request - State: closed - Opened by DamRsn almost 2 years ago - 3 comments

#88 - RFC: Allow use of abstract ModelT

Pull Request - State: closed - Opened by falkTX almost 2 years ago - 9 comments

#87 - Option to use dynamically allocated memory for Conv1D layer state

Pull Request - State: closed - Opened by jatinchowdhury18 almost 2 years ago

#86 - Fix TF model saving and loading for models with plain activation layers

Pull Request - State: closed - Opened by jatinchowdhury18 about 2 years ago - 1 comment

#85 - Model gives erroneous values as prediction

Issue - State: closed - Opened by mikegazzaruso about 2 years ago - 4 comments

#84 - Issue when trying to parse json

Issue - State: closed - Opened by mikegazzaruso about 2 years ago - 3 comments

#83 - Add BatchNorm1D implementation

Pull Request - State: closed - Opened by jatinchowdhury18 about 2 years ago - 1 comment

#82 - Add PReLU layer for all backends

Pull Request - State: closed - Opened by jatinchowdhury18 about 2 years ago - 1 comment

#81 - Update xsimd for Android support.

Pull Request - State: closed - Opened by atsushieno about 2 years ago - 5 comments

#80 - update to the latest xsimd for better performance on Android

Issue - State: closed - Opened by atsushieno about 2 years ago

#79 - Feature/add test pytorch imported models

Pull Request - State: open - Opened by MaxPayne86 about 2 years ago - 5 comments

#78 - XSIMD: add specialization for Conv1D layer with dilation rate = 1

Pull Request - State: closed - Opened by jatinchowdhury18 about 2 years ago - 1 comment

#77 - Conv1D: update XSIMD and STL implementations to use strategy from the Eigen implementation

Pull Request - State: closed - Opened by jatinchowdhury18 about 2 years ago - 1 comment

#76 - Model Loading: use .at() instead of operator[] where appropriate

Pull Request - State: closed - Opened by jatinchowdhury18 about 2 years ago

#75 - parseJson crashes when hidden size different from the one contained in the json model

Issue - State: closed - Opened by MaxPayne86 about 2 years ago - 3 comments

#74 - Improve Conv1D layers (Eigen backend)

Pull Request - State: closed - Opened by stepanmk over 2 years ago - 3 comments

#73 - Improve Conv1D Implementation

Issue - State: closed - Opened by jatinchowdhury18 over 2 years ago

#72 - Quick question on how this performs compared to Pytorch on M1

Issue - State: open - Opened by pgarz over 2 years ago - 3 comments

#71 - Update xsimd

Pull Request - State: closed - Opened by jatinchowdhury18 over 2 years ago - 1 comment

#70 - GELU approx V1

Pull Request - State: closed - Opened by algoravioli over 2 years ago

#69 - Update nlohmann::json to version 3.11.1

Pull Request - State: closed - Opened by jatinchowdhury18 over 2 years ago - 1 comment

#68 - Fix extra inner loops in dense_xsimd for small input size

Pull Request - State: closed - Opened by jatinchowdhury18 over 2 years ago

#67 - Minor improvements for softmax and elu activation functions

Pull Request - State: closed - Opened by jatinchowdhury18 over 2 years ago - 1 comment

#66 - Mark functions as noexcept where appropriate

Pull Request - State: closed - Opened by jatinchowdhury18 over 2 years ago

#65 - Adding sample-rate correction options to templated recurrent layers

Pull Request - State: closed - Opened by jatinchowdhury18 over 2 years ago - 2 comments

#64 - Performance optimizations for XSIMD layers

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#63 - could suport the gpu?

Issue - State: open - Opened by xinsuinizhuan almost 3 years ago - 1 comment

#62 - could support the Dropout layer?

Issue - State: closed - Opened by xinsuinizhuan almost 3 years ago - 1 comment

#61 - Be more explicit about which SIMD instruction sets to compile with

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago

#60 - Use AVX instead of AVX2

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago

#59 - Add example for models with custom layers

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#58 - Add ELu activation layer

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#57 - Smarter enforcement of C++ standard, interfacing include directory

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#56 - Use Eigen::Map for activation layer outputs

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago

#55 - Unexpected Eigen Results

Issue - State: closed - Opened by bhostetler18 almost 3 years ago - 2 comments

#54 - Update XSIMD submodule to version 8.0.5

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#53 - PyTorch state_dict to RTNeural JSON

Issue - State: closed - Opened by ross-ca almost 3 years ago - 9 comments

#52 - Add example Makefile and documentation for building without CMake

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#51 - Benchmarks: add 24x1 LSTM and 1x24 dense

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#50 - Layers support & Input Data issues

Issue - State: open - Opened by vackva almost 3 years ago - 10 comments

#49 - Fix issue with XSIMD needing C++ limits header

Pull Request - State: closed - Opened by jatinchowdhury18 almost 3 years ago - 1 comment

#48 - xsimd numeric includes issue.

Pull Request - State: closed - Opened by rerdavies almost 3 years ago

#47 - <limits> std::numeric_limits include error.

Issue - State: closed - Opened by rerdavies almost 3 years ago - 4 comments

#46 - How do I transform .pt files to json ?

Issue - State: closed - Opened by creativeguitar about 3 years ago - 5 comments

#45 - Add faster LSTM layer with the Eigen backend

Pull Request - State: closed - Opened by jatinchowdhury18 about 3 years ago

#44 - Building out the Accelerate / BNNS backend

Issue - State: closed - Opened by pgarz about 3 years ago - 1 comment

#43 - Use Armadillo as backend

Issue - State: open - Opened by AlessandroProverbioFX about 3 years ago - 3 comments

#42 - Fix 32-byte alignment for Eigen backend

Pull Request - State: closed - Opened by jatinchowdhury18 about 3 years ago - 1 comment

#41 - Improvements to Eigen backend

Pull Request - State: closed - Opened by jatinchowdhury18 about 3 years ago - 1 comment

#40 - Compatible with C++14, CMake 3.1, and run clang-tidy

Pull Request - State: closed - Opened by jatinchowdhury18 about 3 years ago - 1 comment

#39 - SIMD vs memory management performance gain

Issue - State: closed - Opened by zanellia about 3 years ago - 4 comments

#38 - Set up test coverage workflow

Pull Request - State: closed - Opened by jatinchowdhury18 about 3 years ago - 1 comment

#37 - Enable optional compilation with AVX2 SIMD

Pull Request - State: closed - Opened by jatinchowdhury18 about 3 years ago