Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / official-stockfish/nnue-pytorch issues and pull requests

#289 - setting --num-workers has no effect

Issue - State: closed - Opened by protonspring 2 months ago - 1 comment

#288 - Label of train set

Issue - State: open - Opened by pynojs 2 months ago

#287 - conda issue, recent Pytorch support

Issue - State: open - Opened by jdart1 2 months ago

#286 - Easy Train problems w/ Ubuntu 24 (python 3.12)

Issue - State: closed - Opened by protonspring 3 months ago - 23 comments

#285 - no gradient val test

Pull Request - State: open - Opened by AstroBoy1 3 months ago

#284 - Add SFNNv9 Diagram

Pull Request - State: open - Opened by xu-shawn 5 months ago

#283 - Fix Typo In NNUE Diagrams

Pull Request - State: open - Opened by xu-shawn 5 months ago

#282 - use pdep to find nthSetBitIndex

Pull Request - State: closed - Opened by Disservin 6 months ago - 1 comment

#281 - Getting started -- .binpack files missing?

Issue - State: open - Opened by TonyGuil 6 months ago - 2 comments

#280 - Increase L1 size to 3072

Pull Request - State: closed - Opened by linrock 7 months ago - 1 comment

#279 - Quantization losses in the L1 biases of L1 3072

Issue - State: open - Opened by Viren6 8 months ago

#278 - Fix Typo

Pull Request - State: closed - Opened by xu-shawn 8 months ago - 2 comments

#277 - Add formatting

Pull Request - State: open - Opened by Disservin 9 months ago - 3 comments

#276 - Visualizing chess position embeddings

Issue - State: open - Opened by bastiansg 9 months ago - 7 comments

#274 - [MINOR:TYPO] Update run_games.py

Pull Request - State: closed - Opened by cakiki 10 months ago

#273 - Improving documentation

Pull Request - State: closed - Opened by FauziAkram 11 months ago - 1 comment

#272 - Added missing .strategy

Pull Request - State: closed - Opened by Naphthalin 11 months ago

#270 - pytorch alphazero?

Issue - State: closed - Opened by XintianHan about 1 year ago - 1 comment

#269 - Network Testing Issue

Issue - State: closed - Opened by HenryZNNUE about 1 year ago - 2 comments

#268 - Update default testing book to UHO_Lichess_4852_v1.epd.zip

Pull Request - State: closed - Opened by linrock about 1 year ago - 1 comment

#267 - Add SFNNv8 architecture and diagrams

Pull Request - State: closed - Opened by linrock about 1 year ago - 3 comments

#266 - Increase L1 size to 2560 for SFNNv8 nets

Pull Request - State: closed - Opened by linrock about 1 year ago

#265 - Update readme on training.

Pull Request - State: closed - Opened by Sopel97 about 1 year ago

#262 - Update references after repo transfer to official-stockfish

Pull Request - State: closed - Opened by Sopel97 over 1 year ago

#261 - No .nnue files found

Issue - State: closed - Opened by HenryZNNUE over 1 year ago - 4 comments

#260 - train.py stops without errors

Issue - State: closed - Opened by GiacomoPorpiglia over 1 year ago - 1 comment

#259 - Add support for multiple datasets in the trainer. The datasets are in…

Pull Request - State: closed - Opened by Sopel97 over 1 year ago - 10 comments

#258 - Update default network testing book to match new fishtest book

Pull Request - State: closed - Opened by linrock over 1 year ago

#257 - [WIP] Automatic script to fetch current datasets.

Pull Request - State: open - Opened by Sopel97 over 1 year ago - 5 comments

#256 - Add SFNNv7 architecture to docs with updated diagrams

Pull Request - State: closed - Opened by linrock over 1 year ago

#255 - Increase L1 size to 2048 for training SFNNv7 nets

Pull Request - State: closed - Opened by linrock over 1 year ago

#254 - FT optimization utility. Integrate with serialization.

Pull Request - State: closed - Opened by Sopel97 over 1 year ago - 5 comments

#253 - Enable LEB128 compression by default.

Pull Request - State: closed - Opened by Sopel97 over 1 year ago

#252 - Keep description on reserialization.

Pull Request - State: closed - Opened by Sopel97 over 1 year ago

#251 - Support for LEB128 compression of feature transformer parameters.

Pull Request - State: closed - Opened by Sopel97 over 1 year ago - 2 comments

#249 - Add blocked sparse input affine transform approach to docs.

Pull Request - State: closed - Opened by Sopel97 over 1 year ago - 5 comments

#248 - Update table of contents in nnue.md

Pull Request - State: closed - Opened by cj5716 over 1 year ago

#247 - Add SFNNv6 architecture to docs with a new diagram

Pull Request - State: closed - Opened by linrock over 1 year ago - 5 comments

#246 - Increase L1 size to 1536 for new SFNNv6 net

Pull Request - State: closed - Opened by linrock over 1 year ago

#245 - Butterfly Effect Can Cause The Wrong NNUE Evaluation

Issue - State: closed - Opened by ginkgo20 over 1 year ago - 7 comments

#243 - Testing trained model

Issue - State: open - Opened by benliu961 over 1 year ago - 7 comments

#242 - Loss Curves

Issue - State: closed - Opened by benliu961 over 1 year ago - 6 comments

#241 - Error with visualize.py

Issue - State: closed - Opened by benliu961 over 1 year ago - 2 comments
Labels: bug

#240 - Changing epoch size and resulting number of epochs when running train.py

Issue - State: closed - Opened by benliu961 over 1 year ago - 3 comments

#239 - serialize.py not working

Issue - State: closed - Opened by benliu961 over 1 year ago - 9 comments

#238 - Illegal instruction

Issue - State: closed - Opened by benliu961 over 1 year ago - 2 comments

#237 - Pytorch Lightning Version

Issue - State: closed - Opened by benliu961 over 1 year ago

#236 - Wrong input size in SFNNv5 architecture diagram

Issue - State: open - Opened by UniQP over 1 year ago
Labels: documentation

#235 - NNUE documentation: Fix links in table of contents

Pull Request - State: closed - Opened by UniQP over 1 year ago

#234 - serialize with --features=HalfKP no longer works

Issue - State: open - Opened by jdart1 over 1 year ago

#233 - Creating new training data files for new training objective

Issue - State: closed - Opened by evanrussek over 1 year ago - 1 comment

#232 - Creating a new .binpack or .bin dataset without generating it with Stockfish

Issue - State: closed - Opened by benliu961 over 1 year ago - 9 comments

#231 - Improve NNUE documentation for feature factorization.

Pull Request - State: closed - Opened by Sopel97 over 1 year ago

#230 - Improve NNUE documentation

Pull Request - State: closed - Opened by UniQP over 1 year ago

#229 - Licence for Training Data

Issue - State: closed - Opened by ONE-RANDOM-HUMAN over 1 year ago - 1 comment

#228 - Improve nnue.md documentation

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

#227 - Enable using early fen skipping param with easy_train.py

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

#226 - fix another missing param for early_fen_skipping

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

#225 - fix missing parameter

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

#223 - Improve CUDA check for torch

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

#222 - introduce skip early plies parameter

Pull Request - State: closed - Opened by Disservin almost 2 years ago - 4 comments

#221 - Skip all positions with ply less than or equal to 28

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

#220 - Fix 2 images in nnue.md.

Pull Request - State: closed - Opened by Sopel97 almost 2 years ago

#219 - Skip positions with score equal to Stockfish's VALUE_NONE (32002)

Pull Request - State: closed - Opened by Sopel97 almost 2 years ago - 2 comments

#218 - Cannot find data_loader shared library

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

#217 - train_data.bin?

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

#216 - Pretrained checkpoints

Issue - State: closed - Opened by adalsteinnpals almost 2 years ago - 6 comments

#215 - Align column text when writing ordo tables to improve readability

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

#214 - Fix warnings about thread.setDaemon() in python 3.10+

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

#213 - Fix Trainer.root_gpu crash when using pytorch-lightning 1.8+

Pull Request - State: closed - Opened by linrock about 2 years ago - 17 comments

#212 - error on startup: Trainer.root_gpu was deprecated

Issue - State: closed - Opened by jdart1 about 2 years ago - 2 comments

#210 - Fix the incorrect NUM_PT in HalfKAv2_hmFactorized

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

#209 - Fix the incorrect weight scale used in NNUEReader

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

#208 - refine loss function power

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

#207 - model.py issue

Issue - State: closed - Opened by shuangqiao999 over 2 years ago - 2 comments

#206 - Also pass param_index down to the data loader

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

#205 - Refined loss function

Pull Request - State: closed - Opened by vondele over 2 years ago - 8 comments

#202 - Fix binpack reader thread buffer utilization.

Pull Request - State: closed - Opened by Sopel97 over 2 years ago

#200 - [easy_train] Handle trainer/player mismatch on --resume-training

Issue - State: open - Opened by Sopel97 over 2 years ago
Labels: enhancement

#199 - [WIP] Attempt to optimize the position skipping to fit a distribution.

Pull Request - State: open - Opened by Sopel97 over 2 years ago - 1 comment

#198 - How to make the trainer work on CPU only mode now?

Issue - State: closed - Opened by PikaCat-OuO over 2 years ago - 1 comment

#189 - Optional local MSYS2 setup.

Pull Request - State: closed - Opened by Sopel97 over 2 years ago - 4 comments

#149 - Update visualize_multi_hist.py to HalfKAv2_hm-1024x2-8-32-1

Pull Request - State: closed - Opened by Sopel97 about 3 years ago

#141 - Is the training setup for the master network here

Issue - State: closed - Opened by Dboingue over 3 years ago - 2 comments

#139 - Try to split FT into two parts per perspective.

Issue - State: closed - Opened by Sopel97 over 3 years ago

#138 - Deprecation Warnings

Issue - State: closed - Opened by Torom over 3 years ago

#127 - TODO for nnue.md

Issue - State: open - Opened by Sopel97 over 3 years ago
Labels: documentation, enhancement

#100 - Small typo and wording fixes.

Pull Request - State: closed - Opened by Sopel97 over 3 years ago

#99 - Test discord integration

Issue - State: closed - Opened by glinscott over 3 years ago - 1 comment

#98 - Add Table Of Contents for nnue.md

Pull Request - State: closed - Opened by Sopel97 over 3 years ago