[WIP] pd: update paddle doc #4489
Conversation
📝 WalkthroughWalkthroughThis pull request adds comprehensive documentation support for the Paddle deep learning framework in the DeePMD-kit project. The changes span multiple documentation files, including README.md, backend documentation, installation guides, model-specific documentation, and training instructions. The updates consistently integrate Paddle as a new supported backend, adding icons, installation commands, training procedures, and framework-specific details across various sections of the project documentation. Changes
Suggested labels
Suggested reviewers
Possibly related PRs
Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media? 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 5
🧹 Nitpick comments (13)
doc/freeze/freeze.md (1)
38-44: Minor grammar improvement.Consider changing “The output model is called
model.jsonandmodel.pdiparams.” to indicate that there are two distinct files:- The output model is called `model.json` and `model.pdiparams`. + The output files are called `model.json` and `model.pdiparams`.🧰 Tools
🪛 LanguageTool
[uncategorized] ~44-~44: The grammatical number of this noun doesn’t look right. Consider replacing it.
Context: ... where the model is trained. The output model is calledmodel.jsonand `model.pdipa...(AI_EN_LECTOR_REPLACEMENT_NOUN_NUMBER)
🪛 Markdownlint (0.37.0)
41-41: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
doc/env.md (2)
59-64: Consider adding usage examples for PD_JITThe documentation clearly explains the purpose and options but would benefit from practical examples showing how to enable JIT in different scenarios and its impact on performance.
66-72: Add examples and clarify CINN compatibilityWhile the documentation mentions the CINN requirement, it would be helpful to:
- Add examples showing how to check CINN support and enable it
- Provide guidance on when to use CINN (e.g., specific operations or models that benefit from it)
- Document any known limitations or compatibility issues
doc/backend.md (2)
38-46: Enhance Paddle backend documentationThe Paddle backend section would benefit from additional information:
- Platform support (CPU/GPU/ROCm compatibility)
- Performance characteristics compared to other backends
- Migration guide for users switching from TensorFlow or PyTorch
- Known limitations or unsupported features
69-69: Add examples of backend switching commandsConsider adding concrete examples of the
dpcommands for each backend to make it clearer for users:# Train with TensorFlow backend dp train --tf input.json # Train with PyTorch backend dp train --pt input.json # Train with Paddle backend dp train --pd input.jsondoc/train/training.md (1)
29-35: LGTM! Consider adding command output example.The Paddle tab item is well-structured and consistent with other backend documentation. Consider adding an example of the command output to match the documentation style of other backends.
🧰 Tools
🪛 Markdownlint (0.37.0)
32-32: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
doc/train/finetuning.md (2)
255-255: Fix typo in documentation"refering" should be "referring"
-One can check the available model branches in multi-task pre-trained model by refering to the documentation +One can check the available model branches in multi-task pre-trained model by referring to the documentation🧰 Tools
🪛 LanguageTool
[misspelling] ~255-~255: This word is normally spelled as one.
Context: ...n check the available model branches in multi-task pre-trained model by refering to the do...(EN_COMPOUNDS_MULTI_TASK)
208-264: Standardize terminology usageThe document inconsistently uses both hyphenated and non-hyphenated versions of terms:
- "pre-trained" vs "pretrained"
- "multi-task" vs "multitask"
Please standardize the usage throughout the document.
🧰 Tools
🪛 Markdownlint (0.37.0)
214-214: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
221-221: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
228-228: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
251-251: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
258-258: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
🪛 LanguageTool
[uncategorized] ~210-~210: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ilizes a single-task pre-trained model (pretrained.pd) and modifies the energy bias withi...(EN_WORD_COHERENCY)
[misspelling] ~241-~241: This word is normally spelled as one.
Context: ...: {} } ``` #### Fine-tuning from a multi-task pre-trained model Additionally, within...(EN_COMPOUNDS_MULTI_TASK)
[misspelling] ~243-~243: This word is normally spelled as one.
Context: ...bility offered by the framework and the multi-task training process proposed in DPA2 [pape...(EN_COMPOUNDS_MULTI_TASK)
[typographical] ~247-~247: It appears that a comma is missing.
Context: ... the paper above. For fine-tuning using this multitask pre-trained model (`multitask...(BY_FOR_IN_COMMA)
[grammar] ~247-~247: The verb ‘multitask’ is plural. Did you mean: “multitasks”? Did you use a verb instead of a noun?
Context: ...aper above. For fine-tuning using this multitask pre-trained model (`multitask_pretraine...(PLURAL_VERB_AFTER_THIS)
[uncategorized] ~247-~247: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...using this multitask pre-trained model (multitask_pretrained.pd), one can select a specific branch ...(EN_WORD_COHERENCY)
[uncategorized] ~248-~248: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ch (e.g.,CHOOSEN_BRANCH) included inmultitask_pretrained.pdfor fine-tuning with the following ...(EN_WORD_COHERENCY)
[misspelling] ~255-~255: This word is normally spelled as one.
Context: ...n check the available model branches in multi-task pre-trained model by refering to the do...(EN_COMPOUNDS_MULTI_TASK)
doc/train/parallel-training.md (3)
202-202: Fix punctuation and grammarThere are two issues in this line:
- Double period at the end
- Missing article before "sequential sampler"
-> **Note** The underlying dataloader will use a distributed sampler to ensure that each GPU receives batches with different content in parallel mode, which will use sequential sampler in serial mode. In the TensorFlow version, Horovod shuffles the dataset using different random seeds for the same purpose.. +> **Note** The underlying dataloader will use a distributed sampler to ensure that each GPU receives batches with different content in parallel mode, which will use a sequential sampler in serial mode. In the TensorFlow version, Horovod shuffles the dataset using different random seeds for the same purpose.🧰 Tools
🪛 LanguageTool
[uncategorized] ~202-~202: Possible missing article found.
Context: ...ontent in parallel mode, which will use sequential sampler in serial mode. In the TensorFl...(AI_HYDRA_LEO_MISSING_A)
[typographical] ~202-~202: Two consecutive dots
Context: ...ferent random seeds for the same purpose.. ```mermaid flowchart LR subgraph s...(DOUBLE_PUNCTUATION)
193-193: Consider using English documentation URLThe Paddle DDP documentation link is in Chinese. Consider using the English version of the documentation if available.
278-279: Enhance environment variable documentationThe note about
NUM_WORKERScould be more specific about recommended ranges or provide examples of appropriate values.Consider adding more specific guidance:
-If `NUM_WORKERS` is too large, it may cause the program to be terminated by the system; -if it is too small, it may slow down data reading. You can try adjusting it to an appropriate size. +If `NUM_WORKERS` is too large (e.g., > 16), it may cause the program to be terminated by the system; +if it is too small (e.g., < 4), it may slow down data reading. We recommend starting with NUM_WORKERS=8 +and adjusting based on your system's capabilities.doc/install/install-from-source.md (2)
365-366: Fix grammar and improve clarity.The sentence has grammatical issues and can be clearer.
-If you want to use C++ interface of Paddle, you need to compile the Paddle inference library(C++ interface) manually from the [linux-compile-by-make](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/install/compile/linux-compile-by-make.html), then use the `.so` and `.a` files in `Paddle/build/paddle_inference_install_dir/`. +To use the C++ interface of Paddle, you need to manually compile the Paddle inference library following the [Linux compilation guide](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/install/compile/linux-compile-by-make.html). After compilation, use the `.so` and `.a` files from `Paddle/build/paddle_inference_install_dir/`.🧰 Tools
🪛 LanguageTool
[style] ~365-~365: You have already used this phrasing in nearby sentences. Consider replacing it to add variety to your writing.
Context: ...-item} Paddle {{ paddle_icon }} If you want to use C++ interface of Paddle, you need t...(REP_WANT_TO_VB)
441-442: Fix grammar in the assumption statement.The sentence has a grammatical error with the verb "get".
-I assume you have get the Paddle inference library(C++ interface) to `$PADDLE_INFERENCE_DIR`, then execute CMake +I assume you have obtained the Paddle inference library (C++ interface) in `$PADDLE_INFERENCE_DIR`. Execute CMake with:🧰 Tools
🪛 LanguageTool
[grammar] ~441-~441: The verb form ‘get’ does not seem to be suitable in this context.
Context: ...le {{ paddle_icon }} I assume you have get the Paddle inference library(C++ interf...(HAVE_VB)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
doc/_static/paddle.svgis excluded by!**/*.svg
📒 Files selected for processing (16)
README.md(1 hunks)doc/backend.md(3 hunks)doc/conf.py(1 hunks)doc/env.md(1 hunks)doc/freeze/freeze.md(1 hunks)doc/install/easy-install.md(1 hunks)doc/install/install-from-source.md(10 hunks)doc/model/dpa2.md(1 hunks)doc/model/sel.md(1 hunks)doc/model/train-energy.md(1 hunks)doc/model/train-se-atten.md(1 hunks)doc/model/train-se-e2-a.md(1 hunks)doc/train/finetuning.md(2 hunks)doc/train/parallel-training.md(2 hunks)doc/train/tensorboard.md(1 hunks)doc/train/training.md(1 hunks)
✅ Files skipped from review due to trivial changes (2)
- doc/model/train-se-e2-a.md
- doc/model/train-energy.md
🧰 Additional context used
🪛 Markdownlint (0.37.0)
doc/train/training.md
32-32: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
doc/freeze/freeze.md
41-41: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
50-50: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
doc/train/finetuning.md
214-214: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
221-221: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
228-228: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
251-251: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
258-258: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
🪛 LanguageTool
doc/freeze/freeze.md
[uncategorized] ~44-~44: The grammatical number of this noun doesn’t look right. Consider replacing it.
Context: ... where the model is trained. The output model is called model.json and `model.pdipa...
(AI_EN_LECTOR_REPLACEMENT_NOUN_NUMBER)
[misspelling] ~46-~46: This word is normally spelled as one.
Context: ...model.jsonandmodel.pdiparams`. In [multi-task mode](../train/multi-task-training-pd.m...
(EN_COMPOUNDS_MULTI_TASK)
doc/train/parallel-training.md
[uncategorized] ~194-~194: Possible missing comma found.
Context: ...arallel (distributed) mode or in serial mode depending on your execution command. #...
(AI_HYDRA_LEO_MISSING_COMMA)
[misspelling] ~200-~200: This word is normally spelled as one.
Context: ...fines the number of threads involved in multi-threaded loading, which can be modified by setti...
(EN_COMPOUNDS_MULTI_THREADED)
[uncategorized] ~202-~202: Possible missing article found.
Context: ...ontent in parallel mode, which will use sequential sampler in serial mode. In the TensorFl...
(AI_HYDRA_LEO_MISSING_A)
[typographical] ~202-~202: Two consecutive dots
Context: ...ferent random seeds for the same purpose.. ```mermaid flowchart LR subgraph s...
(DOUBLE_PUNCTUATION)
doc/install/install-from-source.md
[style] ~365-~365: You have already used this phrasing in nearby sentences. Consider replacing it to add variety to your writing.
Context: ...-item} Paddle {{ paddle_icon }} If you want to use C++ interface of Paddle, you need t...
(REP_WANT_TO_VB)
[grammar] ~441-~441: The verb form ‘get’ does not seem to be suitable in this context.
Context: ...le {{ paddle_icon }} I assume you have get the Paddle inference library(C++ interf...
(HAVE_VB)
doc/train/finetuning.md
[misspelling] ~202-~202: Did you mean “and”?
Context: ... In Paddle version, we have introduced an updated, more adaptable approach to fin...
(AN_AND)
[uncategorized] ~210-~210: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ilizes a single-task pre-trained model (pretrained.pd) and modifies the energy bias withi...
(EN_WORD_COHERENCY)
[misspelling] ~241-~241: This word is normally spelled as one.
Context: ...: {} } ``` #### Fine-tuning from a multi-task pre-trained model Additionally, within...
(EN_COMPOUNDS_MULTI_TASK)
[misspelling] ~243-~243: This word is normally spelled as one.
Context: ...bility offered by the framework and the multi-task training process proposed in DPA2 [pape...
(EN_COMPOUNDS_MULTI_TASK)
[typographical] ~247-~247: It appears that a comma is missing.
Context: ... the paper above. For fine-tuning using this multitask pre-trained model (`multitask...
(BY_FOR_IN_COMMA)
[grammar] ~247-~247: The verb ‘multitask’ is plural. Did you mean: “multitasks”? Did you use a verb instead of a noun?
Context: ...aper above. For fine-tuning using this multitask pre-trained model (`multitask_pretraine...
(PLURAL_VERB_AFTER_THIS)
[uncategorized] ~247-~247: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...using this multitask pre-trained model (multitask_pretrained.pd), one can select a specific branch ...
(EN_WORD_COHERENCY)
[uncategorized] ~248-~248: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ch (e.g., CHOOSEN_BRANCH) included in multitask_pretrained.pd for fine-tuning with the following ...
(EN_WORD_COHERENCY)
[misspelling] ~255-~255: This word is normally spelled as one.
Context: ...n check the available model branches in multi-task pre-trained model by refering to the do...
(EN_COMPOUNDS_MULTI_TASK)
🔇 Additional comments (10)
doc/conf.py (1)
170-170: Ensure the icon file is present.
The file reference "_static/paddle.svg" is newly introduced. Confirm that it exists and is properly placed under the "_static/" directory; otherwise, documentation builds may fail or show a broken icon link.
Use this script to verify the presence of "paddle.svg" in the repository:
✅ Verification successful
Icon file is present at the expected location
The file paddle.svg exists at doc/_static/paddle.svg, which is the correct location referenced in the configuration.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Search the entire repo for paddle.svg to ensure it exists
fd paddle.svg
Length of output: 38
doc/model/sel.md (1)
35-39: LGTM!
The addition of Paddle is consistent with the existing pattern for other backends. This is well-aligned with the broader documentation updates.
doc/freeze/freeze.md (1)
46-54: LGTM!
The new multi-task instructions for Paddle are clear and consistent with the structure used for other backends.
🧰 Tools
🪛 LanguageTool
[misspelling] ~46-~46: This word is normally spelled as one.
Context: ...model.jsonandmodel.pdiparams`. In [multi-task mode](../train/multi-task-training-pd.m...
(EN_COMPOUNDS_MULTI_TASK)
🪛 Markdownlint (0.37.0)
50-50: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
doc/model/dpa2.md (1)
1-4: No issues found.
Adding the Paddle icon and listing it among supported backends is consistent with the broader changes.
doc/train/tensorboard.md (1)
1-4: Consider adding Paddle-specific TensorBoard instructions
While the Paddle backend has been added to the supported backends list, the document doesn't mention any Paddle-specific considerations or setup requirements for TensorBoard integration. Please clarify:
- Are there any special steps needed to enable TensorBoard with Paddle?
- Are all TensorBoard features (metrics tracking, model graph, etc.) fully supported with Paddle?
README.md (1)
22-22: LGTM! Documentation update is accurate and well-integrated.
The addition of Paddle to the list of supported backends is consistent with the PR objectives and maintains the documentation style.
doc/model/train-se-atten.md (1)
1-1: LGTM: Paddle backend support properly documented
The addition of Paddle icons in the header and supported backends note accurately reflects the new backend support.
Also applies to: 4-4
doc/train/finetuning.md (1)
200-264: LGTM: Paddle implementation properly documented
The Paddle implementation section accurately mirrors the PyTorch implementation, providing clear instructions for both single-task and multi-task fine-tuning scenarios.
🧰 Tools
🪛 Markdownlint (0.37.0)
214-214: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
221-221: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
228-228: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
251-251: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
258-258: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
🪛 LanguageTool
[misspelling] ~202-~202: Did you mean “and”?
Context: ... In Paddle version, we have introduced an updated, more adaptable approach to fin...
(AN_AND)
[uncategorized] ~210-~210: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ilizes a single-task pre-trained model (pretrained.pd) and modifies the energy bias withi...
(EN_WORD_COHERENCY)
[misspelling] ~241-~241: This word is normally spelled as one.
Context: ...: {} } ``` #### Fine-tuning from a multi-task pre-trained model Additionally, within...
(EN_COMPOUNDS_MULTI_TASK)
[misspelling] ~243-~243: This word is normally spelled as one.
Context: ...bility offered by the framework and the multi-task training process proposed in DPA2 [pape...
(EN_COMPOUNDS_MULTI_TASK)
[typographical] ~247-~247: It appears that a comma is missing.
Context: ... the paper above. For fine-tuning using this multitask pre-trained model (`multitask...
(BY_FOR_IN_COMMA)
[grammar] ~247-~247: The verb ‘multitask’ is plural. Did you mean: “multitasks”? Did you use a verb instead of a noun?
Context: ...aper above. For fine-tuning using this multitask pre-trained model (`multitask_pretraine...
(PLURAL_VERB_AFTER_THIS)
[uncategorized] ~247-~247: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...using this multitask pre-trained model (multitask_pretrained.pd), one can select a specific branch ...
(EN_WORD_COHERENCY)
[uncategorized] ~248-~248: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ch (e.g., CHOOSEN_BRANCH) included in multitask_pretrained.pd for fine-tuning with the following ...
(EN_WORD_COHERENCY)
[misspelling] ~255-~255: This word is normally spelled as one.
Context: ...n check the available model branches in multi-task pre-trained model by refering to the do...
(EN_COMPOUNDS_MULTI_TASK)
doc/train/parallel-training.md (1)
191-281: LGTM: Paddle parallel training implementation well documented
The Paddle implementation section effectively documents the DDP setup, providing clear instructions for both single-node and multi-node training scenarios. The dataloader implementation and usage instructions are comprehensive.
🧰 Tools
🪛 LanguageTool
[uncategorized] ~194-~194: Possible missing comma found.
Context: ...arallel (distributed) mode or in serial mode depending on your execution command. #...
(AI_HYDRA_LEO_MISSING_COMMA)
[misspelling] ~200-~200: This word is normally spelled as one.
Context: ...fines the number of threads involved in multi-threaded loading, which can be modified by setti...
(EN_COMPOUNDS_MULTI_THREADED)
[uncategorized] ~202-~202: Possible missing article found.
Context: ...ontent in parallel mode, which will use sequential sampler in serial mode. In the TensorFl...
(AI_HYDRA_LEO_MISSING_A)
[typographical] ~202-~202: Two consecutive dots
Context: ...ferent random seeds for the same purpose.. ```mermaid flowchart LR subgraph s...
(DOUBLE_PUNCTUATION)
doc/install/install-from-source.md (1)
212-218: LGTM! Well-documented environment and CMake variables.
The documentation for environment variables and CMake variables is clear, complete, and follows the same format as other backends.
Also applies to: 233-239, 480-487, 496-503
doc/install/easy-install.md
Outdated
| ```bash | ||
| pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu123/ | ||
| pip install deepmd-kit-cu11 | ||
| ``` |
There was a problem hiding this comment.
Verify package repository URL.
The CUDA 11.8 installation uses the CUDA 12.3 package repository URL. This should be updated to use the correct CUDA version.
-pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu123/
+pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
doc/install/easy-install.md
Outdated
| ```bash | ||
| pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/ | ||
| pip install deepmd-kit | ||
| ``` |
There was a problem hiding this comment.
Fix CPU installation command.
The CPU installation command incorrectly uses the GPU package. For CPU-only installations, use the CPU-specific package.
-pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
+pip install paddlepaddle==3.0.0b2 -i https://www.paddlepaddle.org.cn/whl/linux/cpu/stable.html
doc/install/easy-install.md
Outdated
| ::::::{tab-item} Paddle {{ paddle_icon }} | ||
|
|
||
| :::::{tab-set} | ||
|
|
||
| ::::{tab-item} CUDA 12.3 | ||
|
|
||
| ```bash | ||
| pip install deepmd-kit[paddle] | ||
| ``` | ||
|
|
||
| :::: | ||
|
|
||
| ::::{tab-item} CUDA 11.8 | ||
|
|
||
| ```bash | ||
| pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu123/ | ||
| pip install deepmd-kit-cu11 | ||
| ``` | ||
|
|
||
| :::: | ||
|
|
||
| ::::{tab-item} CPU | ||
|
|
||
| ```bash | ||
| pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/ | ||
| pip install deepmd-kit | ||
| ``` | ||
|
|
||
| :::: | ||
|
|
||
| ::::: | ||
|
|
||
| :::::: |
There was a problem hiding this comment.
💡 Codebase verification
Warning needed: PaddlePaddle 3.0.0b2 is still in beta
The PyPI version history confirms that PaddlePaddle 3.0.0b2 is the latest version and is still in beta status. Since the installation instructions rely on this beta version, we should add explicit warnings about:
- Using beta software in production environments
- Potential API instability and breaking changes
- The lack of a stable 3.0.0 release
Specific locations needing attention:
doc/install/easy-install.md: Add a warning note above the Paddle installation instructions about beta status and stability implications
🔗 Analysis chain
Consider stability implications of beta dependency.
The installation instructions specify PaddlePaddle version 3.0.0b2, which is a beta release. Consider:
- Adding a warning about potential stability issues
- Documenting any known limitations
- Specifying when a stable release is expected
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Check if there are newer versions available
curl -s https://pypi.org/pypi/paddlepaddle/json | jq -r '.releases | keys[]' | sort -V | tail -n 5
Length of output: 136
| [Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz) | ||
|
|
||
| [Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz) |
There was a problem hiding this comment.
💡 Codebase verification
Add last modified dates to pre-compiled library links
The links are actively maintained with recent updates (December 21, 2024). Adding the last modified dates would help users verify the freshness of these pre-compiled libraries:
- Update link text for CUDA 11.8:
[Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz (Last updated: Dec 21, 2024)](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz) - Update link text for CUDA 12.3:
[Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz (Last updated: Dec 21, 2024)](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz)
🔗 Analysis chain
Consider adding version information for pre-compiled libraries.
The links to pre-compiled libraries should include version information or last updated date to help users make informed decisions.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Check if the links are accessible and get last modified date
for url in \
"https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz" \
"https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz"; do
echo "Checking $url"
curl -sI "$url" | grep "Last-Modified"
done
Length of output: 1552
| # cu123 | ||
| pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu123/ | ||
| # cu118 | ||
| pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu118/ | ||
| # cpu | ||
| pip install --pre paddlepaddle -i https://www.paddlepaddle.org.cn/packages/nightly/cpu/ |
There was a problem hiding this comment.
💡 Codebase verification
Use stable version installation commands instead of nightly builds
The repository's easy-install.md already contains stable version installation commands for PaddlePaddle. Replace the nightly build commands with stable version commands:
# cu123
pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu123/
# cu118
pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
# cpu
pip install paddlepaddle==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/
This aligns with the version requirement (3.0 or above) mentioned in backend.md and provides more stable installations compared to nightly builds.
🔗 Analysis chain
Consider providing stable release installation commands.
The current installation commands use nightly builds (--pre flag), which might be unstable. Consider adding commands for installing stable releases or adding a note about potential instability of nightly builds.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Check if stable versions are available
curl -s https://www.paddlepaddle.org.cn/documentation/docs/en/install/pip/linux-pip_en.html | grep -A 5 "pip install paddlepaddle"
Length of output: 132
Script:
#!/bin/bash
# Let's check if there are any mentions of stable PaddlePaddle versions in the repository
rg -i "paddlepaddle.*stable" --type md
# Also check for any version constraints or recommendations
rg -i "paddlepaddle.*version" --type md
# Check if there are any other installation instructions for PaddlePaddle
fd -e md -x grep -l "paddlepaddle" {}
Length of output: 612
Script:
#!/bin/bash
# Let's check the stable version installation commands from easy-install.md
cat doc/install/easy-install.md | grep -A 5 "paddlepaddle.*stable"
# Also check the version requirement mentioned in backend.md
cat doc/backend.md | grep -A 2 -B 2 "Paddle.*version"
Length of output: 989
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## devel #4489 +/- ##
==========================================
- Coverage 84.59% 84.55% -0.04%
==========================================
Files 675 677 +2
Lines 63575 63902 +327
Branches 3486 3487 +1
==========================================
+ Hits 53779 54034 +255
- Misses 8671 8742 +71
- Partials 1125 1126 +1 ☔ View full report in Codecov by Sentry. |
doc/train/parallel-training.md
Outdated
| Currently, parallel training in paddle version is implemented in the form of Paddle Distributed Data Parallelism [DDP](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/guides/06_distributed_training/cluster_quick_start_collective_cn.html). | ||
| DeePMD-kit will decide whether to launch the training in parallel (distributed) mode or in serial mode depending on your execution command. | ||
|
|
||
| ### Dataloader and Dataset |
There was a problem hiding this comment.
Is this section copied from PyTorch? I suggest only maintaining one for the same documentation
There was a problem hiding this comment.
Is this section copied from PyTorch? I suggest only maintaining one for the same documentation
You are right, this is duplicated and removed now.
Co-authored-by: Jinzhe Zeng <[email protected]> Signed-off-by: HydrogenSulfate <[email protected]>
Co-authored-by: Jinzhe Zeng <[email protected]> Signed-off-by: HydrogenSulfate <[email protected]>
…e/deepmd-kit into update_paddle_doc
HydrogenSulfate
force-pushed
the
update_paddle_doc
branch
from
e39ef8b to
af81543
Compare
There was a problem hiding this comment.
Actionable comments posted: 0
♻️ Duplicate comments (1)
doc/install/install-from-source.md (1)
101-106: 🛠️ Refactor suggestionUse stable version installation commands
Replace the nightly build commands with stable version commands for better stability:
-# cu123 -pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu123/ -# cu118 -pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu118/ -# cpu -pip install --pre paddlepaddle -i https://www.paddlepaddle.org.cn/packages/nightly/cpu/ +# cu123 +pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu123/ +# cu118 +pip install paddlepaddle-gpu==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/ +# cpu +pip install paddlepaddle==3.0.0b2 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/
🧹 Nitpick comments (4)
doc/train/finetuning.md (3)
213-215: Format commands according to documentation standards.Remove the
$prefix from commands to align with markdown best practices.Apply this format to all commands:
-$ dp --pd train input.json --finetune pretrained.pd +dp --pd train input.json --finetune pretrained.pdAlso applies to: 220-222, 227-229, 250-252, 257-259
🧰 Tools
🪛 Markdownlint (0.37.0)
214-214: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
247-252: Fix typo in branch name example.Correct the spelling of "CHOOSEN_BRANCH" to "CHOSEN_BRANCH" in the example.
-one can select a specific branch (e.g., `CHOOSEN_BRANCH`) included in `multitask_pretrained.pd` for fine-tuning with the following command: +one can select a specific branch (e.g., `CHOSEN_BRANCH`) included in `multitask_pretrained.pd` for fine-tuning with the following command: -$ dp --pd train input.json --finetune multitask_pretrained.pd --model-branch CHOOSEN_BRANCH +dp --pd train input.json --finetune multitask_pretrained.pd --model-branch CHOSEN_BRANCH🧰 Tools
🪛 Markdownlint (0.37.0)
251-251: null
Dollar signs used before commands without showing output(MD014, commands-show-output)
🪛 LanguageTool
[typographical] ~247-~247: It appears that a comma is missing.
Context: ... the paper above. For fine-tuning using this multitask pre-trained model (`multitask...(BY_FOR_IN_COMMA)
[grammar] ~247-~247: The verb ‘multitask’ is plural. Did you mean: “multitasks”? Did you use a verb instead of a noun?
Context: ...aper above. For fine-tuning using this multitask pre-trained model (`multitask_pretraine...(PLURAL_VERB_AFTER_THIS)
[uncategorized] ~247-~247: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...using this multitask pre-trained model (multitask_pretrained.pd), one can select a specific branch ...(EN_WORD_COHERENCY)
[uncategorized] ~248-~248: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ch (e.g.,CHOOSEN_BRANCH) included inmultitask_pretrained.pdfor fine-tuning with the following ...(EN_WORD_COHERENCY)
254-256: Fix typo in note section.Correct the spelling of "refering" to "referring".
-One can check the available model branches in multi-task pre-trained model by refering to the documentation of the pre-trained model or by using the following command: +One can check the available model branches in multi-task pre-trained model by referring to the documentation of the pre-trained model or by using the following command:🧰 Tools
🪛 LanguageTool
[misspelling] ~255-~255: This word is normally spelled as one.
Context: ...n check the available model branches in multi-task pre-trained model by refering to the do...(EN_COMPOUNDS_MULTI_TASK)
doc/install/install-from-source.md (1)
437-437: Fix grammatical errorThe phrase "have get" is grammatically incorrect.
-I assume you have get the Paddle inference library(C++ interface) +I assume you have obtained the Paddle inference library (C++ interface)🧰 Tools
🪛 LanguageTool
[grammar] ~437-~437: The verb form ‘get’ does not seem to be suitable in this context.
Context: ...le {{ paddle_icon }} I assume you have get the Paddle inference library(C++ interf...(HAVE_VB)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
doc/_static/paddle.svgis excluded by!**/*.svg
📒 Files selected for processing (6)
README.md(1 hunks)doc/install/easy-install.md(1 hunks)doc/install/install-from-source.md(9 hunks)doc/model/dpa2.md(1 hunks)doc/train/finetuning.md(2 hunks)doc/train/parallel-training.md(2 hunks)
🚧 Files skipped from review as they are similar to previous changes (3)
- README.md
- doc/model/dpa2.md
- doc/install/easy-install.md
🧰 Additional context used
🪛 LanguageTool
doc/install/install-from-source.md
[style] ~359-~359: You have already used this phrasing in nearby sentences. Consider replacing it to add variety to your writing.
Context: ...-item} Paddle {{ paddle_icon }} If you want to use C++ interface of Paddle, you need t...
(REP_WANT_TO_VB)
[uncategorized] ~361-~361: Although a hyphen is possible, it is not necessary in a compound modifier in which the first word is an adverb that ends in ‘ly’.
Context: ...rence_install_dir/`. We also provide a weekly-build Paddle C++ library for Linux x86_64 wit...
(HYPHENATED_LY_ADVERB_ADJECTIVE)
[grammar] ~437-~437: The verb form ‘get’ does not seem to be suitable in this context.
Context: ...le {{ paddle_icon }} I assume you have get the Paddle inference library(C++ interf...
(HAVE_VB)
doc/train/finetuning.md
[misspelling] ~202-~202: Did you mean “and”?
Context: ... In Paddle version, we have introduced an updated, more adaptable approach to fin...
(AN_AND)
[uncategorized] ~210-~210: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ilizes a single-task pre-trained model (pretrained.pd) and modifies the energy bias withi...
(EN_WORD_COHERENCY)
[misspelling] ~241-~241: This word is normally spelled as one.
Context: ...: {} } ``` #### Fine-tuning from a multi-task pre-trained model Additionally, within...
(EN_COMPOUNDS_MULTI_TASK)
[misspelling] ~243-~243: This word is normally spelled as one.
Context: ...bility offered by the framework and the multi-task training process proposed in DPA2 [pape...
(EN_COMPOUNDS_MULTI_TASK)
[typographical] ~247-~247: It appears that a comma is missing.
Context: ... the paper above. For fine-tuning using this multitask pre-trained model (`multitask...
(BY_FOR_IN_COMMA)
[grammar] ~247-~247: The verb ‘multitask’ is plural. Did you mean: “multitasks”? Did you use a verb instead of a noun?
Context: ...aper above. For fine-tuning using this multitask pre-trained model (`multitask_pretraine...
(PLURAL_VERB_AFTER_THIS)
[uncategorized] ~247-~247: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...using this multitask pre-trained model (multitask_pretrained.pd), one can select a specific branch ...
(EN_WORD_COHERENCY)
[uncategorized] ~248-~248: Do not mix variants of the same word (‘pretrain’ and ‘pre-train’) within a single text.
Context: ...ch (e.g., CHOOSEN_BRANCH) included in multitask_pretrained.pd for fine-tuning with the following ...
(EN_WORD_COHERENCY)
[misspelling] ~255-~255: This word is normally spelled as one.
Context: ...n check the available model branches in multi-task pre-trained model by refering to the do...
(EN_COMPOUNDS_MULTI_TASK)
🪛 Markdownlint (0.37.0)
doc/train/finetuning.md
214-214: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
221-221: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
228-228: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
251-251: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
258-258: null
Dollar signs used before commands without showing output
(MD014, commands-show-output)
⏰ Context from checks skipped due to timeout of 90000ms (19)
- GitHub Check: Test Python (6, 3.12)
- GitHub Check: Test Python (6, 3.9)
- GitHub Check: Test Python (5, 3.12)
- GitHub Check: Test Python (5, 3.9)
- GitHub Check: Test Python (4, 3.12)
- GitHub Check: Test Python (4, 3.9)
- GitHub Check: Build C++ (clang, clang)
- GitHub Check: Test Python (3, 3.12)
- GitHub Check: Test Python (3, 3.9)
- GitHub Check: Test Python (2, 3.12)
- GitHub Check: Analyze (python)
- GitHub Check: Test Python (2, 3.9)
- GitHub Check: Test C++ (false)
- GitHub Check: Test Python (1, 3.12)
- GitHub Check: Build C library (2.14, >=2.5.0rc0,<2.15, libdeepmd_c_cu11.tar.gz)
- GitHub Check: Test Python (1, 3.9)
- GitHub Check: Build C library (2.18, libdeepmd_c.tar.gz)
- GitHub Check: Analyze (c-cpp)
- GitHub Check: Test C++ (true)
🔇 Additional comments (5)
doc/train/finetuning.md (3)
1-4: LGTM! Header and note section updated correctly.The changes consistently integrate Paddle as a supported backend alongside TensorFlow and PyTorch.
200-203: LGTM! Clear and consistent introduction.The introduction aligns well with the existing PyTorch implementation structure, maintaining documentation consistency.
🧰 Tools
🪛 LanguageTool
[misspelling] ~202-~202: Did you mean “and”?
Context: ... In Paddle version, we have introduced an updated, more adaptable approach to fin...(AN_AND)
264-264: Add multi-task fine-tuning section for Paddle.The documentation is missing the multi-task fine-tuning section for Paddle, which is present in the PyTorch implementation. Please add this section to maintain consistency across backends and provide complete documentation for Paddle users.
Would you like me to help generate the missing multi-task fine-tuning section based on the PyTorch implementation?
doc/train/parallel-training.md (1)
191-225: Documentation for Paddle parallel training looks good!The implementation details, commands, and environment variables are well documented. The examples cover both single-node and multi-node scenarios effectively.
doc/install/install-from-source.md (1)
363-365: Add last modified dates to pre-compiled library linksThe links are actively maintained with recent updates. Adding the last modified dates would help users verify the freshness of these pre-compiled libraries:
-[Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz) +[Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz (Last updated: Dec 21, 2024)](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda118_cudnn860_Trt8531_D1/latest/paddle_inference.tgz) -[Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz) +[Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz (Last updated: Dec 21, 2024)](https://paddle-qa.bj.bcebos.com/paddle-pipeline/GITHUB_Docker_Compile_Test_Cuda123_cudnn900_Trt8616_D1/latest/paddle_inference.tgz)
Add paddle icon svg and update related docs
Summary by CodeRabbit
New Features
Documentation