doc/install/easy-install.md (1)

212-217: CPU install block looks correct; small pip best-practice nit.

Consider using “python -m pip …” to ensure the right interpreter’s pip when users have multiple Pythons.

Apply if you want the consistency:

- pip install paddlepaddle==3.1.1 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/
+ python -m pip install paddlepaddle==3.1.1 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/

doc/train/training.md (1)

32-41: Good clarification; consider scoping the CINN speedup claim.

The wording changes read better. To avoid overpromising, add “depending on model and hardware” next to the “~40%+ speedup” note or cite a benchmark.

-# [experimental] training models with the CINN compiler (~40%+ speedup)
+# [experimental] training models with the CINN compiler (~40%+ speedup, depending on model and hardware)

doc/install/install-from-source.md (1)

102-118: Version bump to Paddle 3.1.1 looks consistent; validate index URLs.

The cu126/cu118/cpu blocks are consistent with easy-install.md. Please verify that 3.1.1 is available on the referenced stable indices, and that the nightly indices are valid, to keep source-install docs in sync with reality.

Also, optional: prefer “python -m pip …” for interpreter-safe installs.

-pip install paddlepaddle-gpu==3.1.1 -i https://www.paddlepaddle.org.cn/packages/stable/cu126/
+python -m pip install paddlepaddle-gpu==3.1.1 -i https://www.paddlepaddle.org.cn/packages/stable/cu126/

deepmd/pd/model/descriptor/repformers.py (1)

69-71: Fix minor grammar and spacing in the NotImplementedError message.

There is an extra space before the period after “model ”, and we can slightly tighten the phrasing.

Apply:

-            "The 'border_op' operator is unavailable because the custom Paddle OP library was not built when freezing the model .\n"
-            "To install 'border_op', run: python source/op/pd/setup.py install\n"
-            "For more information, please refer to the DPA3 documentation."
+            "The 'border_op' operator is unavailable because the custom Paddle OP library was not built when freezing the model.\n"
+            "To install 'border_op', run: python source/op/pd/setup.py install\n"
+            "For more information, please refer to the DPA3 documentation."

deepmd/pd/model/descriptor/repflows.py (1)

72-74: Fix minor grammar and spacing in the NotImplementedError message.

Same extra space before the period and minor phrasing nit as in repformers.py.

Apply:

-            "The 'border_op' operator is unavailable because the custom Paddle OP library was not built when freezing the model .\n"
-            "To install 'border_op', run: python source/op/pd/setup.py install\n"
-            "For more information, please refer to the DPA3 documentation."
+            "The 'border_op' operator is unavailable because the custom Paddle OP library was not built when freezing the model.\n"
+            "To install 'border_op', run: python source/op/pd/setup.py install\n"
+            "For more information, please refer to the DPA3 documentation."

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro

💡 Knowledge Base configuration:

• MCP integration is disabled by default for public repositories
• Jira integration is disabled by default for public repositories
• Linear integration is disabled by default for public repositories

You can enable these settings in your CodeRabbit configuration.

Learnt from: njzjz
PR: deepmodeling/deepmd-kit#4884
File: .github/workflows/test_cuda.yml:46-46
Timestamp: 2025-08-14T07:11:51.333Z
Learning: As of PyTorch 2.8 (August 2025), the default wheel on PyPI installed by `pip install torch` is CPU-only. CUDA-enabled wheels are available on PyPI for Linux x86 and Windows x86 platforms, but require explicit specification via index URLs or variant-aware installers. For CUDA support, use `--index-url https://download.pytorch.org/whl/cu126` (or appropriate CUDA version).

doc/model/dpa3.md (4)

1-1: Supported backends note is inconsistent with the header (missing Paddle).

Header shows Paddle support, but the “Supported backends” note (Line 4) omits Paddle. Please update the note for accuracy.

Proposed change (adjust the note text on Line 4):

**Supported backends**: PyTorch {{ pytorch_icon }}, JAX {{ jax_icon }}, Paddle {{ paddle_icon }}, DP {{ dpmodel_icon }}

---

43-43: Minor wording tweak for the section title.

“Requirements for installation from source code” reads better.

-## Requirements of installation from source code {{ pytorch_icon }} {{ paddle_icon }}
+## Requirements for installation from source code {{ pytorch_icon }} {{ paddle_icon }}

---

47-60: Consider adding explicit install commands for the PyTorch OP path.

You mention setting DP_ENABLE_PYTORCH=1 during installation but don’t provide an example. Adding a short, copy-pastable snippet helps users avoid misconfigurations.

Suggested insertion after Line 54:

 The customized OP library for the Python interface can be installed by setting environment variable {envvar}`DP_ENABLE_PYTORCH` to `1` during installation.
+
+For example:
+
+```sh
+# from the repository root
+export DP_ENABLE_PYTORCH=1
+python -m pip install -v .
+```

---

64-69: Replace deprecated “setup.py install” with pip.

Directly invoking setup.py is deprecated and can lead to environment inconsistencies. Prefer pip to handle builds and metadata.

 ```sh
 cd deepmd-kit/source/op/pd
-python setup.py install
+python -m pip install -v .

Optional: if build isolation causes issues with compiler flags or local deps, document the fallback:

python -m pip install -v --no-build-isolation .

</blockquote></details>

</blockquote></details>

<details>
<summary>📜 Review details</summary>

**Configuration used: CodeRabbit UI**
**Review profile: CHILL**
**Plan: Pro**

**💡 Knowledge Base configuration:**

- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories

You can enable these sources in your CodeRabbit configuration.

<details>
<summary>📥 Commits</summary>

Reviewing files that changed from the base of the PR and between 696941227c0885f3cfc0561a762b6561935a6c7e and f03083049200ec24a343c5d8627b2e13db33e136.

</details>

<details>
<summary>📒 Files selected for processing (1)</summary>

* `doc/model/dpa3.md` (3 hunks)

</details>

<details>
<summary>⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (29)</summary>

* GitHub Check: Build C library (2.14, >=2.5.0,<2.15, libdeepmd_c_cu11.tar.gz)
* GitHub Check: Build C library (2.18, libdeepmd_c.tar.gz)
* GitHub Check: Build wheels for cp310-manylinux_aarch64
* GitHub Check: Build wheels for cp311-macosx_x86_64
* GitHub Check: Build wheels for cp311-win_amd64
* GitHub Check: Build wheels for cp311-macosx_arm64
* GitHub Check: Build wheels for cp311-manylinux_x86_64
* GitHub Check: Build wheels for cp311-manylinux_x86_64
* GitHub Check: Build C++ (clang, clang)
* GitHub Check: Build C++ (cuda120, cuda)
* GitHub Check: Build C++ (rocm, rocm)
* GitHub Check: Build C++ (cpu, cpu)
* GitHub Check: Build C++ (cuda, cuda)
* GitHub Check: Test C++ (false)
* GitHub Check: Test Python (3, 3.9)
* GitHub Check: Test C++ (true)
* GitHub Check: Test Python (3, 3.12)
* GitHub Check: Test Python (5, 3.12)
* GitHub Check: Test Python (6, 3.12)
* GitHub Check: Test Python (6, 3.9)
* GitHub Check: Test Python (5, 3.9)
* GitHub Check: Test Python (1, 3.12)
* GitHub Check: Test Python (4, 3.12)
* GitHub Check: Test Python (2, 3.9)
* GitHub Check: Test Python (4, 3.9)
* GitHub Check: Test Python (2, 3.12)
* GitHub Check: Test Python (1, 3.9)
* GitHub Check: Analyze (python)
* GitHub Check: Analyze (c-cpp)

</details>

<details>
<summary>🔇 Additional comments (1)</summary><blockquote>

<details>
<summary>doc/model/dpa3.md (1)</summary>

`71-74`: **LGTM on MPI/CUDA guidance; mirrors PyTorch tab.**

The MPI/CUDA recommendations are clear and consistent with the PyTorch tab. No action required.

</details>

</blockquote></details>

</details>

<!-- This is an auto-generated comment by CodeRabbit for review status -->