This is an automated email from the ASF dual-hosted git repository.

ruihangl pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git


The following commit(s) were added to refs/heads/main by this push:
     new 4fc4450412 [FIX] Fix T.sblock due to concurrent merge (#18695)
4fc4450412 is described below

commit 4fc4450412eeb138edc4bd83ee17d1cefbf144ff
Author: Tianqi Chen <[email protected]>
AuthorDate: Thu Jan 29 10:24:25 2026 -0500

    [FIX] Fix T.sblock due to concurrent merge (#18695)
    
    This PR fixes a CI issue due to concurrent merge of a refactor.
---
 docs/deep_dive/tensor_ir/learning.rst                | 2 +-
 tests/python/relax/test_transform_legalize_ops_nn.py | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/docs/deep_dive/tensor_ir/learning.rst 
b/docs/deep_dive/tensor_ir/learning.rst
index ace7ebe9de..8a02274431 100644
--- a/docs/deep_dive/tensor_ir/learning.rst
+++ b/docs/deep_dive/tensor_ir/learning.rst
@@ -137,7 +137,7 @@ Loop Iterations
 Computational Block
 ~~~~~~~~~~~~~~~~~~~
 A significant distinction lies in computational statements:
-**TensorIR incorporates an additional construct termed** ``T.block``.
+**TensorIR incorporates an additional construct termed** ``T.sblock``.
 
 .. code:: python
 
diff --git a/tests/python/relax/test_transform_legalize_ops_nn.py 
b/tests/python/relax/test_transform_legalize_ops_nn.py
index 3121c31ac8..f1ba7105ae 100644
--- a/tests/python/relax/test_transform_legalize_ops_nn.py
+++ b/tests/python/relax/test_transform_legalize_ops_nn.py
@@ -3918,7 +3918,7 @@ def test_batch_flatten():
         def reshape(x: T.Buffer((T.int64(2), T.int64(3), T.int64(4), 
T.int64(5)), "float32"), T_reshape: T.Buffer((T.int64(2), T.int64(60)), 
"float32")):
             T.func_attr({"tir.noalias": True})
             for ax0, ax1 in T.grid(T.int64(2), T.int64(60)):
-                with T.block("T_reshape"):
+                with T.sblock("T_reshape"):
                     v_ax0, v_ax1 = T.axis.remap("SS", [ax0, ax1])
                     T.reads(x[(v_ax1 // T.int64(60) + v_ax0) % T.int64(2), 
v_ax1 % T.int64(60) // T.int64(20), v_ax1 % T.int64(20) // T.int64(5), v_ax1 % 
T.int64(5)])
                     T.writes(T_reshape[v_ax0, v_ax1])

Reply via email to