From e9d19aca39bebc8d61d328bc7aa521629e82156d Mon Sep 17 00:00:00 2001
From: shuw <shuw@nvidia.com>
Date: Wed, 4 Sep 2024 11:52:44 -0700
Subject: [PATCH] Remove unnecessary warnings.

---
 flax/linen/fp8_ops.py | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/flax/linen/fp8_ops.py b/flax/linen/fp8_ops.py
index 598a03c758..767ad45a17 100644
--- a/flax/linen/fp8_ops.py
+++ b/flax/linen/fp8_ops.py
@@ -299,12 +299,6 @@ def q_dot_dq_impl(
     preferred_element_type,
     is_training
 ):
-  if precision != None or preferred_element_type != None:
-    warnings.warn(
-      "The function dot_general_with_precision will set the "
-      "precision/preferred_element_type and disregard any provided "
-      "values."
-    )
   new_lhs_scale, new_lhs_amax_history = quantize_and_update(
     lhs,
     jnp.float8_e4m3fn,