From 29ab3a6b1433a4100e62f54b1382f4dd1028c06f Mon Sep 17 00:00:00 2001 From: Tim Dettmers Date: Tue, 11 Apr 2023 09:26:52 -0700 Subject: [PATCH] Updated change log. --- CHANGELOG.md | 14 ++++++++++++++ setup.py | 2 +- 2 files changed, 15 insertions(+), 1 deletion(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index ac239de..66578d5 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -201,3 +201,17 @@ Features: Improvements: - Improved logging for the CUDA detection mechanism. + +### 0.38.0 + +#### 8-bit Lion, Load/Store 8-bit layers + +Features: + - Support for 32 and 8-bit Lion has been added. Thank you @lucidrains + - Support for serialization of Linear8bitLt layers (LLM.int8()). This allows to store and load 8-bit weights directly from the HuggingFace Hub. Thank you @myrab + +Bug fixes: + - Fixed a bug where some bitsandbytes methods failed in a model-parallel setup on multiple GPUs. Thank you @tonylins + +Deprecated: + - Devices with compute capability 3.0 (GTX 700s, K10) and 3.2 (Tegra K1, Jetson TK1) are now deprecated and support will be removed in 0.39.0. diff --git a/setup.py b/setup.py index a0bbc7f..b023c0b 100644 --- a/setup.py +++ b/setup.py @@ -18,7 +18,7 @@ def read(fname): setup( name=f"bitsandbytes", - version=f"0.37.2", + version=f"0.38.0", author="Tim Dettmers", author_email="dettmers@cs.washington.edu", description="8-bit optimizers and matrix multiplication routines.",