Merge branch 'main' into main

Author: Svetlana Karslioglu

intermediate_source/dynamic_quantization_bert_tutorial.rst

@@ -253,8 +253,6 @@ model before and after the dynamic quantization.
         torch.manual_seed(seed)
     set_seed(42)
 
-    # Initialize a global random number generator
-    global_rng = random.Random()
 
 
 2.2 Load the fine-tuned BERT model
@@ -526,20 +524,9 @@ We can serialize and save the quantized model for the future use using
 
 .. code:: python
 
-    def ids_tensor(shape, vocab_size, rng=None, name=None):
+    def ids_tensor(shape, vocab_size):
         #  Creates a random int32 tensor of the shape within the vocab size
-        if rng is None:
-            rng = global_rng
-
-        total_dims = 1
-        for dim in shape:
-            total_dims *= dim
-
-        values = []
-        for _ in range(total_dims):
-            values.append(rng.randint(0, vocab_size - 1))
-
-        return torch.tensor(data=values, dtype=torch.long, device='cpu').view(shape).contiguous()
+        return torch.randint(0, vocab_size, shape=shape, dtype=torch.int, device='cpu')
 
     input_ids = ids_tensor([8, 128], 2)
     token_type_ids = ids_tensor([8, 128], 2)

intermediate_source/torch_compile_tutorial.py

@@ -69,7 +69,7 @@
 
 def foo(x, y):
     a = torch.sin(x)
-    b = torch.cos(x)
+    b = torch.cos(y)
     return a + b
 opt_foo1 = torch.compile(foo)
 print(opt_foo1(torch.randn(10, 10), torch.randn(10, 10)))
@@ -80,7 +80,7 @@ def foo(x, y):
 @torch.compile
 def opt_foo2(x, y):
     a = torch.sin(x)
-    b = torch.cos(x)
+    b = torch.cos(y)
     return a + b
 print(opt_foo2(torch.randn(10, 10), torch.randn(10, 10)))
 

prototype_source/graph_mode_dynamic_bert_tutorial.rst

@@ -60,22 +60,9 @@ Once all the necesessary packages are downloaded and installed we setup the code
     from torch.quantization import per_channel_dynamic_qconfig
     from torch.quantization import quantize_dynamic_jit
 
-    global_rng = random.Random()
-
-    def ids_tensor(shape, vocab_size, rng=None, name=None):
+    def ids_tensor(shape, vocab_size):
         #  Creates a random int32 tensor of the shape within the vocab size
-        if rng is None:
-            rng = global_rng
-
-        total_dims = 1
-        for dim in shape:
-            total_dims *= dim
-
-        values = []
-        for _ in range(total_dims):
-            values.append(rng.randint(0, vocab_size - 1))
-
-        return torch.tensor(data=values, dtype=torch.long, device='cpu').view(shape).contiguous()
+        return torch.randint(0, vocab_size, shape=shape, dtype=torch.int, device='cpu')
 
     # Setup logging
     logger = logging.getLogger(__name__)