Merge pull request #20687 from rajasekharporeddy:test_branch4

PiperOrigin-RevId: 623877846

Author: jax authors

docs/developer.md

@@ -206,7 +206,7 @@ python build/build.py --configure_only
 You may pass additional options to `build.py` to configure the build; see the
 `jaxlib` build documentation for details.
 
-By default the Bazel build runs the JAX tests using `jaxlib` built form source.
+By default the Bazel build runs the JAX tests using `jaxlib` built from source.
 To run JAX tests, run:
 
 ```

docs/faq.rst

@@ -413,7 +413,7 @@ speed of code using JAX:
    use 32-bit dtypes in NumPy or enable 64-bit dtypes in JAX (see
    `Double (64 bit) precision`_) for a fair comparison.
 4. **Transferring data between CPUs and accelerators takes time.** If you only
-   want to measure the how long it takes to evaluate a function, you may want to
+   want to measure how long it takes to evaluate a function, you may want to
    transfer data to the device on which you want to run it first (see
    :ref:`faq-data-placement`).
 

docs/gpu_performance_tips.md

@@ -69,7 +69,7 @@ training on Nvidia GPUs](https://github.com/NVIDIA/JAX-Toolbox/blob/main/rosetta
   The default value is False.
 * **--xla_gpu_enable_pipelined_collectives** When using pipeline parallelism,
   this flag enables overlapping the (i+1)-th layer weight `AllGather` with the
-  i-th layer computation. It also enables enable overlapping (i+1)-th layer
+  i-th layer computation. It also enables overlapping (i+1)-th layer
   weight `Reduce`/`ReduceScatter` with i-th layer's computation. The default
   value is False. **There are some bugs when this flag is turned on.**
 * **--xla_gpu_collective_permute_decomposer_threshold** This flag is useful when

docs/jax-101/06-parallelism.ipynb

@@ -326,7 +326,7 @@
     "id": "iuHqht-OYqca"
    },
    "source": [
-    "The outputs of the inner `jax.pmap(convolve)` never left their devices when being fed into the outer `jax.pmap(convolve)`."
+    "The outputs of the inner `jax.pmap(convolve)` have never left their devices when being fed into the outer `jax.pmap(convolve)`."
    ]
   },
   {

docs/jax-101/06-parallelism.md

@@ -125,7 +125,7 @@ jax.pmap(convolve)(xs, jax.pmap(convolve)(xs, ws))
 
 +++ {"id": "iuHqht-OYqca"}
 
-The outputs of the inner `jax.pmap(convolve)` never left their devices when being fed into the outer `jax.pmap(convolve)`.
+The outputs of the inner `jax.pmap(convolve)` have never left their devices when being fed into the outer `jax.pmap(convolve)`.
 
 +++ {"id": "vEFAJXN2q3dV"}
 

docs/notebooks/Common_Gotchas_in_JAX.ipynb

@@ -920,7 +920,7 @@
     "id": "ORMVVGZJgSVi"
    },
    "source": [
-    "Underneath the hood, numpy uses the [Mersenne Twister](https://en.wikipedia.org/wiki/Mersenne_Twister) PRNG to power its pseudorandom functions.  The PRNG has a period of $2^{19937}-1$ and at any point can be described by __624 32bit unsigned ints__ and a __position__ indicating how much of this  \"entropy\" has been used up."
+    "Underneath the hood, numpy uses the [Mersenne Twister](https://en.wikipedia.org/wiki/Mersenne_Twister) PRNG to power its pseudorandom functions.  The PRNG has a period of $2^{19937}-1$ and at any point can be described by __624 32-bit unsigned ints__ and a __position__ indicating how much of this  \"entropy\" has been used up."
    ]
   },
   {
@@ -2135,7 +2135,7 @@
     "\n",
     "There are a few ways to do this:\n",
     "\n",
-    "1. You can enable 64bit mode by setting the environment variable `JAX_ENABLE_X64=True`.\n",
+    "1. You can enable 64-bit mode by setting the environment variable `JAX_ENABLE_X64=True`.\n",
     "\n",
     "2. You can manually set the `jax_enable_x64` configuration flag at startup:\n",
     "\n",

docs/notebooks/Common_Gotchas_in_JAX.md

@@ -407,7 +407,7 @@ print(np.random.random())
 
 +++ {"id": "ORMVVGZJgSVi"}
 
-Underneath the hood, numpy uses the [Mersenne Twister](https://en.wikipedia.org/wiki/Mersenne_Twister) PRNG to power its pseudorandom functions.  The PRNG has a period of $2^{19937}-1$ and at any point can be described by __624 32bit unsigned ints__ and a __position__ indicating how much of this  "entropy" has been used up.
+Underneath the hood, numpy uses the [Mersenne Twister](https://en.wikipedia.org/wiki/Mersenne_Twister) PRNG to power its pseudorandom functions.  The PRNG has a period of $2^{19937}-1$ and at any point can be described by __624 32-bit unsigned ints__ and a __position__ indicating how much of this  "entropy" has been used up.
 
 ```{code-cell} ipython3
 :id: 7Pyp2ajzfPO2
@@ -1081,7 +1081,7 @@ To use double-precision numbers, you need to set the `jax_enable_x64` configurat
 
 There are a few ways to do this:
 
-1. You can enable 64bit mode by setting the environment variable `JAX_ENABLE_X64=True`.
+1. You can enable 64-bit mode by setting the environment variable `JAX_ENABLE_X64=True`.
 
 2. You can manually set the `jax_enable_x64` configuration flag at startup: