docs: more sentence case

docs/notebooks/Common_Gotchas_in_JAX.ipynb

@@ -258,7 +258,7 @@
     "id": "oBdKtkVW8Lha"
    },
    "source": [
-    "## 🔪 In-Place Updates"
+    "## 🔪 In-place updates"
    ]
   },
   {
@@ -533,7 +533,7 @@
     "id": "oZ_jE2WAypdL"
    },
    "source": [
-    "## 🔪 Out-of-Bounds Indexing"
+    "## 🔪 Out-of-bounds indexing"
    ]
   },
   {
@@ -868,7 +868,7 @@
     "id": "MUycRNh6e50W"
    },
    "source": [
-    "## 🔪 Random Numbers"
+    "## 🔪 Random numbers"
    ]
   },
   {
@@ -888,7 +888,7 @@
     "id": "Qikt9pPW9L5K"
    },
    "source": [
-    "### RNGs and State\n",
+    "### RNGs and state\n",
     "You're used to _stateful_ pseudorandom number generators (PRNGs) from numpy and other libraries, which helpfully hide a lot of details under the hood to give you a ready fountain of pseudorandomness:"
    ]
   },
@@ -1183,7 +1183,7 @@
     "id": "rg4CpMZ8c3ri"
    },
    "source": [
-    "## 🔪 Control Flow"
+    "## 🔪 Control flow"
    ]
   },
   {
@@ -1192,7 +1192,7 @@
     "id": "izLTvT24dAq0"
    },
    "source": [
-    "### ✔ python control_flow + autodiff ✔\n",
+    "### ✔ Python control_flow + autodiff ✔\n",
     "\n",
     "If you just want to apply `grad` to your python functions, you can use regular python control-flow constructs with no problems, as if you were using [Autograd](https://github.com/hips/autograd) (or Pytorch or TF Eager)."
    ]
@@ -1231,7 +1231,7 @@
     "id": "hIfPT7WMmZ2H"
    },
    "source": [
-    "### python control flow + JIT\n",
+    "### Python control flow + JIT\n",
     "\n",
     "Using control flow with `jit` is more complicated, and by default it has more constraints.\n",
     "\n",
@@ -1791,7 +1791,7 @@
     "id": "OxLsZUyRt_kF"
    },
    "source": [
-    "## 🔪 Dynamic Shapes"
+    "## 🔪 Dynamic shapes"
    ]
   },
   {
@@ -2194,7 +2194,7 @@
     "id": "WAHjmL0E2XwO"
    },
    "source": [
-    "## 🔪 Miscellaneous Divergences from NumPy\n",
+    "## 🔪 Miscellaneous divergences from NumPy\n",
     "\n",
     "While `jax.numpy` makes every attempt to replicate the behavior of numpy's API, there do exist corner cases where the behaviors differ.\n",
     "Many such cases are discussed in detail in the sections above; here we list several other known places where the APIs diverge.\n",

docs/notebooks/Common_Gotchas_in_JAX.md

@@ -158,7 +158,7 @@ iter_operand = iter(range(10))
 
 +++ {"id": "oBdKtkVW8Lha"}
 
-## 🔪 In-Place Updates
+## 🔪 In-place updates
 
 +++ {"id": "JffAqnEW4JEb"}
 
@@ -268,7 +268,7 @@ For more details on indexed array updates, see the [documentation for the `.at`
 
 +++ {"id": "oZ_jE2WAypdL"}
 
-## 🔪 Out-of-Bounds Indexing
+## 🔪 Out-of-bounds indexing
 
 +++ {"id": "btRFwEVzypdN"}
 
@@ -385,7 +385,7 @@ jnp.sum(jnp.array(x))
 
 +++ {"id": "MUycRNh6e50W"}
 
-## 🔪 Random Numbers
+## 🔪 Random numbers
 
 +++ {"id": "O8vvaVt3MRG2"}
 
@@ -395,7 +395,7 @@ jnp.sum(jnp.array(x))
 
 +++ {"id": "Qikt9pPW9L5K"}
 
-### RNGs and State
+### RNGs and state
 You're used to _stateful_ pseudorandom number generators (PRNGs) from numpy and other libraries, which helpfully hide a lot of details under the hood to give you a ready fountain of pseudorandomness:
 
 ```{code-cell} ipython3
@@ -538,11 +538,11 @@ for subkey in subkeys:
 
 +++ {"id": "rg4CpMZ8c3ri"}
 
-## 🔪 Control Flow
+## 🔪 Control flow
 
 +++ {"id": "izLTvT24dAq0"}
 
-### ✔ python control_flow + autodiff ✔
+### ✔ Python control_flow + autodiff ✔
 
 If you just want to apply `grad` to your python functions, you can use regular python control-flow constructs with no problems, as if you were using [Autograd](https://github.com/hips/autograd) (or Pytorch or TF Eager).
 
@@ -562,7 +562,7 @@ print(grad(f)(4.))  # ok!
 
 +++ {"id": "hIfPT7WMmZ2H"}
 
-### python control flow + JIT
+### Python control flow + JIT
 
 Using control flow with `jit` is more complicated, and by default it has more constraints.
 
@@ -865,7 +865,7 @@ $\ast$ = argument-<b>value</b>-independent loop condition - unrolls the loop
 
 +++ {"id": "OxLsZUyRt_kF"}
 
-## 🔪 Dynamic Shapes
+## 🔪 Dynamic shapes
 
 +++ {"id": "1tKXcAMduDR1"}
 
@@ -1130,7 +1130,7 @@ x.dtype # --> dtype('float64')
 
 +++ {"id": "WAHjmL0E2XwO"}
 
-## 🔪 Miscellaneous Divergences from NumPy
+## 🔪 Miscellaneous divergences from NumPy
 
 While `jax.numpy` makes every attempt to replicate the behavior of numpy's API, there do exist corner cases where the behaviors differ.
 Many such cases are discussed in detail in the sections above; here we list several other known places where the APIs diverge.