add a deprecation warning for float8 delayed and static scaling

Summary:

As titled, the complexity tax for these features is high and there no
known real use cases, as the community is overwhelmingly using dynamic
scaling. So, IMO we should deprecate this.

Test Plan: CI

Reviewers:

Subscribers:

Tasks:

Tags:

ghstack-source-id: 2fc91dbf0556728475624bb70f51553d3fa7042b
ghstack-comment-id: 2641358141
Pull Request resolved: #1681

torchao/float8/README.md

@@ -65,6 +65,8 @@ for _ in range(10):
 
 ## float8 linear with delayed scaling
 
+:warning: <em>We plan to deprecate delayed scaling in a future release, see https://github.com/pytorch/ao/issues/1680 for more details.</em>
+
 This is theoretically the most performant recipe as it minimizes memory reads.
 
 ```python

torchao/float8/config.py

@@ -304,6 +304,16 @@ def __post_init__(self):
                 "When using FSDP, it's recommended to enable config.force_recompute_fp8_weight_in_bwd."
             )
 
+        # Future deprecation warning for delayed scaling
+        if (
+            self.cast_config_input.scaling_type != ScalingType.DYNAMIC
+            or self.cast_config_weight.scaling_type != ScalingType.DYNAMIC
+            or self.cast_config_grad_output.scaling_type != ScalingType.DYNAMIC
+        ):
+            logger.warning(
+                "Note: delayed and static scaling will be deprecated in a future release of torchao. Please see https://github.com/pytorch/ao/issues/1680 for more details."
+            )
+
 
 # Pre-made recipes for common configurations
 # TODO(future PR): go through a round of design on this, and eventually expose