modify rev version

Author: roll-away

graph_net/analysis_util.py

@@ -406,12 +406,12 @@ def print_stat_info(
             )
 
             expected_s = (
-                alpha ** lambda_
+                alpha**lambda_
                 * beta ** (lambda_ * eta * negative_speedup_penalty)
                 * fpdb ** (1 - lambda_)
             )
             expected_es = (
-                alpha ** lambda_
+                alpha**lambda_
                 * beta ** (lambda_ * eta * negative_speedup_penalty)
                 * gamma ** (1 - lambda_)
             )

graph_net/samples_statistics.py

@@ -221,7 +221,7 @@ def calculate_gamma(
         if tolerance < errno_tolerance
     )
 
-    return b ** pi_sum
+    return b**pi_sum
 
 
 def calculate_s_t_from_aggregated(
@@ -249,7 +249,7 @@ def calculate_s_t_from_aggregated(
         S(t) value calculated from aggregated parameters
     """
     return (
-        alpha ** lambda_
+        alpha**lambda_
         * beta ** (lambda_ * eta * negative_speedup_penalty)
         * b ** (1 - lambda_)
     )
@@ -280,7 +280,7 @@ def calculate_es_t_from_aggregated(
         ES(t) value calculated from aggregated parameters
     """
     return (
-        alpha ** lambda_
+        alpha**lambda_
         * beta ** (lambda_ * eta * negative_speedup_penalty)
         * gamma ** (1 - lambda_)
     )

graph_net/test_compiler_util.py

@@ -280,11 +280,11 @@ def check_equal(args, expected_out, compiled_out, cmp_equal_func):
 
 def tolerance_generator(t):
     # for float16
-    yield 10 ** (t * 3 / 5), 10 ** t
+    yield 10 ** (t * 3 / 5), 10**t
     # for bfloat16
-    yield 10 ** (t * 1.796 / 5), 10 ** t
+    yield 10 ** (t * 1.796 / 5), 10**t
     # yield float32
-    yield 10 ** (t * 5.886 / 5), 10 ** t
+    yield 10 ** (t * 5.886 / 5), 10**t
     # yield float64
     yield 10 ** (t * 7 / 5), 10 ** (t * 7 / 5)
 

samples/mmseg/DNLNet_R101/model.py

@@ -3875,7 +3875,7 @@ def forward(
             256, dtype=torch.float32, device=device(type="cuda", index=0)
         )
         tensor_1 = torch.tensor(0.5, device=device(type="cuda", index=0))
-        pow_1 = tensor ** tensor_1
+        pow_1 = tensor**tensor_1
         tensor = tensor_1 = None
         pairwise_weight /= pow_1
         pairwise_weight_1 = pairwise_weight

samples/mmseg/DNLNet_R50/model.py

@@ -2107,7 +2107,7 @@ def forward(
             256, dtype=torch.float32, device=device(type="cuda", index=0)
         )
         tensor_1 = torch.tensor(0.5, device=device(type="cuda", index=0))
-        pow_1 = tensor ** tensor_1
+        pow_1 = tensor**tensor_1
         tensor = tensor_1 = None
         pairwise_weight /= pow_1
         pairwise_weight_1 = pairwise_weight

samples/mmseg/Mask2Former_R50/model.py

@@ -441,7 +441,7 @@ def forward(
         floordiv = None
         truediv_2 = mul_2 / 128
         mul_2 = None
-        dim_t_1 = 10000 ** truediv_2
+        dim_t_1 = 10000**truediv_2
         truediv_2 = None
         getitem_3 = x_embed_1[
             (
@@ -571,7 +571,7 @@ def forward(
         floordiv_1 = None
         truediv_7 = mul_5 / 128
         mul_5 = None
-        dim_t_3 = 10000 ** truediv_7
+        dim_t_3 = 10000**truediv_7
         truediv_7 = None
         getitem_12 = x_embed_3[
             (
@@ -702,7 +702,7 @@ def forward(
         floordiv_2 = None
         truediv_12 = mul_8 / 128
         mul_8 = None
-        dim_t_5 = 10000 ** truediv_12
+        dim_t_5 = 10000**truediv_12
         truediv_12 = None
         getitem_21 = x_embed_5[
             (