[docs] Fix examples/ paths in the docs after skyrl-train -> skyrl migration#1269
[docs] Fix examples/ paths in the docs after skyrl-train -> skyrl migration#1269
examples/ paths in the docs after skyrl-train -> skyrl migration#1269Conversation
Signed-off-by: SumanthRH <sumanthrh99@gmail.com>
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request updates file paths in the documentation to reflect the migration of examples from examples/ to examples/train/. The changes are mostly correct and cover a large number of files. However, I've found a couple of instances where Python module paths in uv run commands were not updated, which could lead to confusion and errors for users following the documentation. I've added comments on the relevant files pointing out these missed updates.
| The training configuration uses GRPO with colocated training and generation. Key parameters include: | ||
|
|
||
| **Training configuration** (from `examples/llm_as_a_judge/run_llm_judge.sh`): | ||
| **Training configuration** (from `examples/train/llm_as_a_judge/run_llm_judge.sh`): |
There was a problem hiding this comment.
Thanks for updating the path in this heading. However, in the following code block, the uv run command on line 97 still refers to the old module path examples.llm_as_a_judge.main_llm_judge. This should be updated to examples.train.llm_as_a_judge.main_llm_judge to reflect the new directory structure.
| ### Step 2: Config knobs to tune for fully async training | ||
|
|
||
| Following [examples/fully_async/fully_async_run_gsm8k.sh](https://github.com/NovaSky-AI/SkyRL/blob/main/examples/train/fully_async/fully_async_run_gsm8k.sh), update the training configuration to use your new entrypoint `main_async.py`: | ||
| Following [examples/train/fully_async/fully_async_run_gsm8k.sh](https://github.com/NovaSky-AI/SkyRL/blob/main/examples/train/fully_async/fully_async_run_gsm8k.sh), update the training configuration to use your new entrypoint `main_async.py`: |
There was a problem hiding this comment.
While you've correctly updated the path to fully_async_run_gsm8k.sh, the example command in the code block below (line 79) still uses the old Python module path. The path examples.fully_async.main_async should be updated to examples.train.fully_async.main_async to match the new directory structure.
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
Signed-off-by: SumanthRH <sumanthrh99@gmail.com>
Signed-off-by: SumanthRH <sumanthrh99@gmail.com>
1 participant
Uh oh!
There was an error while loading. Please reload this page.