yanlinjizi
  • Joined on Mar 31, 2023
Public Activity Repositories Starred Repositories
3
 Following
0
 Followers
0
 Achievement Badge
1
Loading Heatmap…

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • e6ccde226d Update AMD targets (#1832) * Update README.md * Update CMakeLists.txt

4 days ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 96ce09353d Bump dev version
  • 26f0c7a9c0 Release 0.49.0
  • 3bff01d50b Fix: Python 3.14 compatibility with PyTorch 2.9 (#1831) * Fix: Python 3.14 / torch.compile compatibility * Skip torch.compile test on Python 3.14 and torch < 2.10 (not supported) * Format
  • c6640545cf README: fix badge
  • 4d127852d9 Add release for DGX Spark cuda121 (#1829) * build for DGX Spark cuda121 * Apply suggestion from @matthewdouglas --------- Co-authored-by: Matthew Douglas <38992547+matthewdouglas@users.noreply.github.com>
  • Compare 6 commits »

6 days ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 5ea4afe8a6 CPU: workaround avx512 4bit dequantize accuracy issue for large blocksize (#1828)

1 week ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 4d19869189 CUDA/ROCm: Remove dead code (#1827) * CUDA/ROCm: Remove dead code * more cleanup
  • 3c71007afc Hf kernel (#1814) * enable hf kernel Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * add kernels dep Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * update tests Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * optional for kernels Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * update kernel Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix format Signed-off-by: jiqing-feng <jiqing.feng@intel.com> --------- Signed-off-by: jiqing-feng <jiqing.feng@intel.com>
  • Compare 2 commits »

1 week ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

1 week ago

yanlinjizi synced commits to master at yanlinjizi/StarrySky from mirror

1 week ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 190d3e2250 ROCm: Add gfx1150/gfx1151 to build targets (#1822)

1 week ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • f6854da7f4 CUDA 13: aggressive compression of binary size (#1820)
  • 61c359db89 ROCm: reduce size of builds (#1819)
  • 4dc08e642c Enable publishing of macOS wheel (#1818) * Enable publishing of macOS wheel * Update macOS target to 14.0+ * Update doc
  • 177494cb1d Cleanup: remove FastBinarySearch (#1817)
  • 54477ddcdb Update README (#1816) Updates the README with the following changes: * Update test workflow status badge to point to new workflow * Indicate AVX512BF16 optimized path for CPU in support table * Indicate macOS CPU support * Indicate slow macOS MPS support * Indicate AMD CDNA4/RDNA4 supported targets
  • Compare 5 commits »

2 weeks ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 6aa9619397 Cpu fused kernel (#1804) * add template to support more dtypes Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * update cmake list Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix compile cpu Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * make different dtype works Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * use bf16 on CPU Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix state2 dtype Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * remove torch Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm torch Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable float to bf16 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm dequantizeBlockwise4bitCpu Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix check Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable dequant 4bit kernel Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix dequantize Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * test Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * change input param Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix input param Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * spliut 8bit and 4bit Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix input params Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix input params Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable dequant4bit Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix reverse Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix dequant 4bit fallback path Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix fp4 dequant Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm _Float16 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * tmp codes Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable gemv Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * change to 4bit dequant Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix def Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix type Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix absmax dtype Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix type Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix compile and type Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable gemv Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix shape Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix lib name Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * debug Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * update Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable gemv 4bit bf16 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable avx512 check Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix check Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix endif Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix format Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix format Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix def Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix position Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix format Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm duplicated func Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm useless code comments Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix out shape Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix comments Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * add reverse format Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * check avx512bf15 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix has_avx512bf16 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix tests Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix absmax shhape Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix compile Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix tests Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix test_gemv Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * disable binsearch Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix lint Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix save Signed-off-by: jiqing-feng <jiqing.feng@intel.com> --------- Signed-off-by: jiqing-feng <jiqing.feng@intel.com>

2 weeks ago

yanlinjizi synced commits to master at yanlinjizi/StarrySky from mirror

3 weeks ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • bd028b886e Remove old nightly workflow (#1812)
  • b5a49aef68 CI: Run tests on PRs, refactor nightly test workflow (#1811) * CI: Run tests on PRs, refactor nightly test workflow * Ensure artifact names are unique * Simplify * Fix for Windows * Fix for Windows * Temp rename * Temp * Update
  • Compare 2 commits »

3 weeks ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 221b4b4e13 Replace NULL with nullptr in pythonInterface.cpp (#1809)

4 weeks ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 76f45feee9 CI: Enable tests on Linux x86-64 with CUDA 13 (#1808)
  • cf68bf5f80 ROCm: Add build for ROCm 7.1 (#1807) * ROCm: Add build for ROCm 7.1 * Use newer CMake from PyPI for ROCm build
  • 5c0a0a9f03 CUDA: Drop compilation compatibility with Maxwell (#1806) * CUDA: Drop Maxwell compatibility * Update docs
  • 8f5d139134 Enable more tests on AMD for warp size 32 (#1805) * Enable even more unit tests for warp size 32 * Revert comment chagnes from previous PR for consistency
  • Compare 4 commits »

1 month ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • 3f9f6f3cbc add support for 64 block size on 32 warp size supported amd gpus (#1748) * add support for 64 block size on 32 warp size supported amd gpus * uncomment 64 block size support in csrc * only enable 64 block size support on architectures with 32 warp size * use BNB_WARP_SIZE instead of warpSize in ops.hip * Reuse BNB_WARP_SIZE macro * Remove unused WARP_SIZE definitions * remove unused import * Apply suggestion from @matthewdouglas * Apply suggestion from @matthewdouglas * Apply suggestion from @matthewdouglas --------- Co-authored-by: sstamenk <strahinja.stamenkovic@amd.com> Co-authored-by: Matthew Douglas <38992547+matthewdouglas@users.noreply.github.com>
  • d1c2b0d004 CI: skip rebuilding CPU lib when building/installing wheels (#1803) * CI: skip rebuilding CPU lib when building/installing wheels * CI: more verbosity when installing bitsandbytes
  • bcdc4def4f fix build error: "no case matching constant switch condition" (#1802) * fix switch error * Apply suggestion from @matthewdouglas --------- Co-authored-by: Matthew Douglas <38992547+matthewdouglas@users.noreply.github.com>
  • 9e589a29b3 Cpu C++ kernel (#1789) * add template to support more dtypes Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * update cmake list Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix compile cpu Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * make different dtype works Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * use bf16 on CPU Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix state2 dtype Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * remove torch Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm torch Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable float to bf16 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm dequantizeBlockwise4bitCpu Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix check Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable dequant 4bit kernel Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix dequantize Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * test Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * change input param Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix input param Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * spliut 8bit and 4bit Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix input params Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix input params Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable dequant4bit Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix reverse Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix dequant 4bit fallback path Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix fp4 dequant Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * rm _Float16 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix cmake check Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix lint Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix datatypr Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix include Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix typo Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix include Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * add runtime check for avx512 Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * enable windows cpu build Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix format Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * Fix some tests * Use larger shape for test --------- Signed-off-by: jiqing-feng <jiqing.feng@intel.com> Co-authored-by: Matthew Douglas <38992547+matthewdouglas@users.noreply.github.com>
  • Compare 4 commits »

1 month ago

yanlinjizi synced commits to master at yanlinjizi/StarrySky from mirror

1 month ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

1 month ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

1 month ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

1 month ago

yanlinjizi synced commits to xpu-windows-build at yanlinjizi/bitsandbytes from mirror

  • 333fdb6437 Update pre-commit to allow crlf line ending on windows batch files

1 month ago

yanlinjizi synced commits to main at yanlinjizi/bitsandbytes from mirror

  • fd9934c3e9 XPU: Add Windows build for SYCL kernels (#1787) * XPU: Build SYCL kernels on Windows * Use shell for Windows XPU build * Update XPU Windows script * Update XPU Windows script * Update XPU Windows script * Update XPU Windows script * Update pre-commit to allow crlf line ending on windows batch files

1 month ago

Please read the following content carefully:

Dear OpenI User

Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.

For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》

Agree and continue
Disagree, exit
Baidu
map