Bump version to 0.2.1

flash_attn/losses/cross_entropy_parallel.py

@@ -14,7 +14,7 @@ from apex.transformer.tensor_parallel.utils import VocabUtility
 
 # `all_gather_into_tensor` and `reduce_scatter_tensor` are new placeholders for
 # `_all_gather_base` and `_reduce_scatter_base`. They require the most recent
-# version of PyTorch. The following 4 lines are for backward comparability with
+# version of PyTorch. The following 4 lines are for backward compatibility with
 # older PyTorch.
 if "all_gather_into_tensor" not in dir(torch.distributed):
     torch.distributed.all_gather_into_tensor = torch.distributed._all_gather_base

setup.py

@@ -152,7 +152,7 @@ ext_modules.append(
 
 setup(
     name="flash_attn",
-    version="0.2.0",
+    version="0.2.1",
     packages=find_packages(
         exclude=("build", "csrc", "include", "tests", "dist", "docs", "benchmarks", "flash_attn.egg-info",)
     ),