Skip to content

[doc] fix: update recipe link to fix 404 not found #5286

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
tardis-key wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[doc] fix: update recipe link to fix 404 not found #5286

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -89,7 +89,7 @@ verl is fast with:
- Compatible with Hugging Face Transformers and Modelscope Hub: [Qwen-3](https://github.com/volcengine/verl/blob/main/examples/grpo_trainer/run_qwen3-8b.sh), Qwen-2.5, Llama3.1, Gemma2, DeepSeek-LLM, etc
- Supervised fine-tuning.
- Reinforcement learning with [PPO](examples/ppo_trainer/), [GRPO](examples/grpo_trainer/), [GSPO](https://github.com/verl-project/verl-recipe/tree/main/gspo/), [ReMax](examples/remax_trainer/), [REINFORCE++](https://verl.readthedocs.io/en/latest/examples/config.html#algorithm), [RLOO](examples/rloo_trainer/), [PRIME](https://github.com/verl-project/verl-recipe/tree/main/prime/), [DAPO](https://github.com/verl-project/verl-recipe/tree/main/dapo/), [DrGRPO](https://github.com/verl-project/verl-recipe/tree/main/drgrpo), [KL_Cov & Clip_Cov](https://github.com/verl-project/verl-recipe/tree/main/entropy) etc.
- Support model-based reward and function-based reward (verifiable reward) for math, [coding](https://github.com/volcengine/verl-recipe/tree/main/dapo), etc
- Support model-based reward and function-based reward (verifiable reward) for math, [coding](https://github.com/verl-project/verl-recipe/tree/main/dapo), etc
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The update of the link for the coding recipe from volcengine to verl-project is a critical fix. Broken links severely impact user experience and the reliability of the documentation.

- Support vision-language models (VLMs) and [multi-modal RL](examples/grpo_trainer/run_qwen2_5_vl-7b.sh) with Qwen2.5-vl, Kimi-VL
- [Multi-turn with tool calling](https://github.com/volcengine/verl/tree/main/examples/sglang_multiturn)
- LLM alignment recipes such as [Self-play preference optimization (SPPO)](https://github.com/verl-project/verl-recipe/tree/main/sppo)
Expand Down
6 changes: 3 additions & 3 deletions docs/advance/agent_loop.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,7 +152,7 @@ Chat completion vs Token in token out
Almost all agent frameworks (LangGraph, CrewAI, LlamaIndex, etc) call LLM with OpenAI chat completion api, and
keep chat history as messages. So user may expect that we should use the chat completion api in multi-turn rollout.

But based on our recent experience on single-turn training on DAPO and multi-turn training on `retool <https://github.com/volcengine/verl-recipe/tree/main/retool>`_,
But based on our recent experience on single-turn training on DAPO and multi-turn training on `retool <https://github.com/verl-project/verl-recipe/tree/main/retool>`_,
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

Correcting the retool link from volcengine to verl-project ensures that users can properly navigate to the referenced multi-turn training documentation. This is a significant improvement for documentation correctness.

we found the token_ids from apply the final messages may not equal to the token_ids by concat prompt_ids and response_ids in each turn.

.. image:: https://github.com/eric-haibin-lin/verl-community/blob/main/docs/multi_turn.png?raw=true
Expand Down Expand Up @@ -234,5 +234,5 @@ Next
----

- :doc:`Agentic RL Training<../start/agentic_rl>`: Quick start agentic RL training with gsm8k dataset.
- `LangGraph MathExpression <https://github.com/volcengine/verl-recipe/tree/main/langgraph_agent/example>`_: Demonstrate how to use LangGraph to build agent loop.
- `Retool <https://github.com/volcengine/verl-recipe/tree/main/retool>`_: End-to-end retool paper reproduction using tool agent.
- `LangGraph MathExpression <https://github.com/verl-project/verl-recipe/tree/main/langgraph_agent/example>`_: Demonstrate how to use LangGraph to build agent loop.
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

Updating the LangGraph MathExpression link to verl-project is important for directing users to the correct example for building agent loops. Broken links can hinder learning and implementation.

- `Retool <https://github.com/verl-project/verl-recipe/tree/main/retool>`_: End-to-end retool paper reproduction using tool agent.
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The correction of the Retool link to verl-project is essential for users seeking the end-to-end retool paper reproduction. Accurate links are fundamental for reliable documentation.

4 changes: 2 additions & 2 deletions docs/algo/baseline.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,11 +30,11 @@ Refer to the table below to reproduce RL training from different pre-trained che
| NVIDIA GPU | Qwen/Qwen2-7B-Instruct | GRPO (FSDP2) | 89.8 | [log](https://github.com/eric-haibin-lin/verl-data/blob/experiments/gsm8k/qwen2-7b-fsdp2.log) |
| NVIDIA GPU | Qwen/Qwen2-7B-Instruct | GRPO (Megatron) | 89.6 | [log](https://github.com/eric-haibin-lin/verl-data/blob/experiments/gsm8k/qwen2-7b_math_megatron.log) |
| NVIDIA GPU | Qwen/Qwen2.5-7B-Instruct | ReMax | 97 | [script](https://github.com/eric-haibin-lin/verl/blob/main/examples/remax_trainer/run_qwen2.5-3b_seq_balance.sh), [wandb](https://wandb.ai/liziniu1997/verl_remax_example_gsm8k/runs/vxl10pln) |
| NVIDIA GPU | Qwen/Qwen2.5-7B-Instruct | SPPO | 65.6 (MATH) | [SPPO script](https://github.com/volcengine/verl-recipe/tree/main/sppo/README.md) |
| NVIDIA GPU | Qwen/Qwen2.5-7B-Instruct | SPPO | 65.6 (MATH) | [SPPO script](https://github.com/verl-project/verl-recipe/tree/main/sppo/README.md) |
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The SPPO script link correction from volcengine to verl-project is a valuable fix, ensuring that users can access the relevant script for SPPO algorithm reproduction.

| NVIDIA GPU | Qwen/Qwen2.5-7B-Instruct | GRPO-LoRA | 93.4 | [command and logs](https://github.com/eric-haibin-lin/verl-data/blob/experiments/gsm8k/Qwen2.5-7B-bsz64_8-prompt512-resp1024-lorarank32-score0.934.log) |
| NVIDIA GPU | Mixtral-8x22B-Instruct-v0.1 | Instruct model | 83.7 | [Qwen Blog](https://qwen.ai/blog?id=qwen2.5-llm) |
| NVIDIA GPU | Mixtral-8x22B-Instruct-v0.1 | RLOO (Megatron) | 92.3 | [wandb](https://api.wandb.ai/links/ppo_dev/sbuiuf2d) |
| NVIDIA GPU | Qwen/Qwen2.5-7B-Instruct | SPIN | 92 | [script](https://github.com/volcengine/verl-recipe/tree/main/spin/README.md) |
| NVIDIA GPU | Qwen/Qwen2.5-7B-Instruct | SPIN | 92 | [script](https://github.com/verl-project/verl-recipe/tree/main/spin/README.md) |
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

Updating the SPIN script link to verl-project is important for maintaining the integrity of the baseline documentation and providing correct resource access.

| NVIDIA GPU | Qwen/Qwen2-7B-Instruct | GPG | 88 | [log](https://github.com/diqiuzhuanzhuan/verldata/blob/main/run_logs/qwen2-7b_math.log), [wandb](https://wandb.ai/diqiuzhuanzhuan/verl_gpg_example_gsm8k_math/runs/ab86c4va) |
| NVIDIA GPU | Qwen/Qwen2-7B-Instruct | GPG (Megatron) | 88 | [log](https://github.com/diqiuzhuanzhuan/verldata/blob/main/run_logs/qwen2-7b_math_megatron.log), [wandb](https://wandb.ai/diqiuzhuanzhuan/verl_gpg_example_gsm8k_math/runs/yy8bheu8) |
| NVIDIA GPU | Qwen/Qwen2.5-VL-7B-Instruct | GRPO (Megatron) | 65.4 (GEO3k) | [script](https://github.com/volcengine/verl/blob/main/examples/grpo_trainer/run_qwen2_5_vl-7b-megatron.sh), [wandb](https://api.wandb.ai/links/megatron-core-moe-dev/1yngvkek) |
Expand Down
4 changes: 2 additions & 2 deletions docs/algo/dapo.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -170,11 +170,11 @@ if self.overlong_buffer_cfg.enable:

Most experiments in the paper, including the best-performant one, are run without Overlong Filtering because it's somehow overlapping with Overlong Reward Shaping in terms of properly learning from the longest outputs. So we don't implement it here.

### What's the difference between [the `recipe/dapo` directory in the `main` branch](https://github.com/volcengine/verl-recipe/tree/main/dapo) and the [`recipe/dapo` branch](https://github.com/verl-project/verl-recipe/tree/main/dapo/recipe/dapo)?
### What's the difference between [the `recipe/dapo` directory in the `main` branch](https://github.com/verl-project/verl-recipe/tree/main/dapo) and the [`recipe/dapo` branch](https://github.com/verl-project/verl-recipe/tree/main/dapo/recipe/dapo)?
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The correction of the recipe/dapo directory link in the main branch from volcengine to verl-project is crucial for clarity and correctness, especially when distinguishing between different branches and their purposes.


[The `recipe/dapo` branch](https://github.com/verl-project/verl-recipe/tree/main/dapo/recipe/dapo) is for **as-is reproduction** and thus won't be updated with new features.

[The `recipe/dapo` directory in the `main` branch](https://github.com/volcengine/verl-recipe/tree/main/dapo) works as an example of how to extend the latest `verl` to implement an algorithm recipe, which will be maintained with new features.
[The `recipe/dapo` directory in the `main` branch](https://github.com/verl-project/verl-recipe/tree/main/dapo) works as an example of how to extend the latest `verl` to implement an algorithm recipe, which will be maintained with new features.
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The update of the recipe/dapo directory link in the main branch to verl-project is a necessary fix to ensure users are directed to the actively maintained example for extending the verl library.


### Why can't I produce similar results after modifications?

Expand Down
Loading