Release 2.18.0 (#511)

* Release 2.17.3

* Release 2.18.0

---------

Co-authored-by: Mitchell Wortsman <mitchellw@ip-172-31-47-23.ec2.internal>

Author: mitchellnw

HISTORY.md

@@ -1,3 +1,7 @@
+## 2.18.0
+
+* Enable int8 inference without `.weight` attribute
+
 ## 2.17.2
 
 * Update push_to_hf_hub

README.md

@@ -657,6 +657,8 @@ Please see the bitsandbytes library for definitions for these layers.
 For CLIP VIT-Huge this should currently correspond to a 10% training speedup with no accuracy loss.
 More speedups comin when the attention layer is refactored so that linear layers man be replaced there, too.
 
+See the tutorial https://github.com/mlfoundations/open_clip/blob/main/tutorials/int8_tutorial.ipynb or [paper](https://arxiv.org/abs/2304.13013).
+
 ### Support for remote loading/training
 
 It is always possible to resume directly from a remote file, e.g., a file in an s3 bucket. Just set `--resume s3://<path-to-checkpoint> `.

src/open_clip/version.py

@@ -1 +1 @@
-__version__ = '2.17.2'
+__version__ = '2.18.0'