docs: fix broken links in 0.1.0 documentation

chenopis · cursoragent · chenopis · commit cd16bedb3a2b · 2026-02-24T14:41:39.000-08:00
Fix 21 broken links across 7 doc files identified via broken link audit
of published docs at docs.nvidia.com/nemo/automodel/0.1.0/.

Fixes:
- fp8_training.md: fix GitHub org (NVIDIA/NeMo-Automodel → NVIDIA-NeMo/Automodel)
  and Llama subdir (llama/ → llama3_1/)
- peft.md, sft.md: update recipe link (finetune.py → train_ft.py) and
  example script path (examples/llm/ → examples/llm_finetune/)
- gemma3-3n.md: insert missing components/ in dataset URLs
- pipelining.md: fix 7 example script paths and 1 YAML config path
- installation.md: fix example script path
- README.md: fix llama subdir (llama/ → llama3_2/)

Co-authored-by: Cursor &lt;cursoragent@cursor.com&gt;
Signed-off-by: Andrew Chen &lt;andrewch@nvidia.com&gt;
diff --git a/README.md b/README.md
@@ -75,7 +75,7 @@ To get started quickly, NeMo AutoModel provides a collection of ready-to-use rec
 **And more**: Check out more [LLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune) and [VLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune) examples! Any causal LM on Hugging Face Hub can be used with the base recipe template!
 
 ### Run a Recipe
-To run a NeMo AutoModel recipe, you need a recipe script (e.g., [LLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/finetune.py), [VLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/finetune.py)) and a YAML config file (e.g., [LLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/llama/llama3_2_1b_squad.yaml), [VLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/gemma3/gemma3_vl_4b_cord_v2_peft.yaml)):
+To run a NeMo AutoModel recipe, you need a recipe script (e.g., [LLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/finetune.py), [VLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/finetune.py)) and a YAML config file (e.g., [LLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/llama3_2/llama3_2_1b_squad.yaml), [VLM](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/gemma3/gemma3_vl_4b_cord_v2_peft.yaml)):
 ```
 # Command invocation format:
 uv run <recipe_script_path> --config <yaml_config_path>
diff --git a/docs/guides/fp8_training.md b/docs/guides/fp8_training.md
@@ -100,12 +100,12 @@ FP8 quantization provides measurable performance improvements while maintaining
 ## Ready-to-Use Recipes
 We provide FP8 training configs for popular models:
 
-- **Llama**: [Llama 3.1 8B](https://github.com/NVIDIA/NeMo-Automodel/blob/main/examples/llm_finetune/llama/llama3_1_8b_hellaswag_fp8.yaml)
-- **Mistral**: [Mistral 7B](https://github.com/NVIDIA/NeMo-Automodel/blob/main/examples/llm_finetune/mistral/mistral_7b_hellaswag_fp8.yaml), [Mistral Nemo 2407](https://github.com/NVIDIA/NeMo-Automodel/blob/main/examples/llm_finetune/mistral/mistral_nemo_2407_hellaswag_fp8.yaml) 
-- **Qwen**: [Qwen 2.5 7B](https://github.com/NVIDIA/NeMo-Automodel/blob/main/examples/llm_finetune/qwen/qwen2_5_7b_hellaswag_fp8.yaml)
-- **Phi**: [Phi 4](https://github.com/NVIDIA/NeMo-Automodel/blob/main/examples/llm_finetune/phi/phi_4_hellaswag_fp8.yaml)
+- **Llama**: [Llama 3.1 8B](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/llama3_1/llama3_1_8b_hellaswag_fp8.yaml)
+- **Mistral**: [Mistral 7B](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/mistral/mistral_7b_hellaswag_fp8.yaml), [Mistral Nemo 2407](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/mistral/mistral_nemo_2407_hellaswag_fp8.yaml) 
+- **Qwen**: [Qwen 2.5 7B](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/qwen/qwen2_5_7b_hellaswag_fp8.yaml)
+- **Phi**: [Phi 4](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/llm_finetune/phi/phi_4_hellaswag_fp8.yaml)
 
-Check out our [examples directory](https://github.com/NVIDIA/NeMo-Automodel/tree/main/examples/llm_finetune/) for more recipes and configurations.
+Check out our [examples directory](https://github.com/NVIDIA-NeMo/Automodel/tree/main/examples/llm_finetune/) for more recipes and configurations.
 
 To run any of these FP8 training recipes, use the following command:
 
diff --git a/docs/guides/installation.md b/docs/guides/installation.md
@@ -107,7 +107,7 @@ docker run --gpus all -it --rm \
   nvcr.io/nvidia/nemo:25.07 /bin/bash -c "\
     cd /workspace/Automodel && \           # Enter the mounted repo
     pip install -e . && \                  # Install Automodel in editable mode
-    python3 examples/llm/finetune.py" # Run a usage example
+    python3 examples/llm_finetune/finetune.py" # Run a usage example
 ```
 :::{note}
 The above `docker` command uses the volume `-v` option to mount the local `Automodel` directory
diff --git a/docs/guides/llm/peft.md b/docs/guides/llm/peft.md
@@ -301,10 +301,10 @@ automodel finetune llm -c peft_guide.yaml
 
 ### Invoke the Recipe Script Directly
 
-Alternatively, you can run the recipe [script](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/recipes/llm/finetune.py) directly using [torchrun](https://docs.pytorch.org/docs/stable/elastic/run.html), as shown below.
+Alternatively, you can run the recipe [script](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/recipes/llm/train_ft.py) directly using [torchrun](https://docs.pytorch.org/docs/stable/elastic/run.html), as shown below.
 
 ``` bash
-torchrun --nproc-per-node=8 examples/llm/finetune.py --config peft_guide.yaml
+torchrun --nproc-per-node=8 examples/llm_finetune/finetune.py --config peft_guide.yaml
 ```
 
 ### Sample Output
diff --git a/docs/guides/llm/sft.md b/docs/guides/llm/sft.md
@@ -29,7 +29,7 @@ vLLM.
 
 <!-- In addition to this user guide, you can also explore our Quickstart,
 which features a [standalone python3
-recipe](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/recipes/llm/finetune.py),
+recipe](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/recipes/llm/train_ft.py),
 offering hands-on demonstrations for quickly getting started with NeMo Automodel. -->
 
 ## Run SFT with NeMo Automodel
@@ -258,10 +258,10 @@ where `finetune` is name the name of the recipe file (excluding the `.py` extens
 
 ### Invoke the Recipe Script Directly
 
-Alternatively, you can run the recipe [script](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/recipes/llm/finetune.py) directly using [torchrun](https://docs.pytorch.org/docs/stable/elastic/run.html), as shown below.
+Alternatively, you can run the recipe [script](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/recipes/llm/train_ft.py) directly using [torchrun](https://docs.pytorch.org/docs/stable/elastic/run.html), as shown below.
 
 ``` bash
-torchrun --nproc-per-node=8 examples/llm/finetune.py --config sft_guide.yaml
+torchrun --nproc-per-node=8 examples/llm_finetune/finetune.py --config sft_guide.yaml
 ```
 
 ### Sample Output
diff --git a/docs/guides/omni/gemma3-3n.md b/docs/guides/omni/gemma3-3n.md
@@ -92,7 +92,7 @@ The default collate function:
 ### Preprocess Custom Datasets
 
 When using a custom dataset with a model whose Hugging Face `AutoProcessor` supports the `apply_chat_template` method, you'll need to convert your data into the Hugging Face message list format expected by the `apply_chat_template`.
-We provide [examples](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/datasets/vlm/datasets.py) demonstrating how to perform this conversion.
+We provide [examples](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/components/datasets/vlm/datasets.py) demonstrating how to perform this conversion.
 
 Some models, such as [Qwen2.5 VL](https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct), have specific preprocessing requirements and require custom collate functions. For instance, Qwen2.5-VL uses the `qwen_vl_utils.process_vision_info` function to process images:
 
@@ -119,7 +119,7 @@ dataloader:
     _target_: nemo_automodel.datasets.vlm.collate_fns.qwen2_5_collate_fn
 ```
 
-We provide [example custom collate functions](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/datasets/vlm/collate_fns.py) that you can use as references for your implementation.
+We provide [example custom collate functions](https://github.com/NVIDIA-NeMo/Automodel/blob/main/nemo_automodel/components/datasets/vlm/collate_fns.py) that you can use as references for your implementation.
 
 ## Run the Fine-Tune Script
 
diff --git a/docs/guides/pipelining.md b/docs/guides/pipelining.md
@@ -102,7 +102,7 @@ For a complete training example:
 
 ```bash
 # Run fine-tuning with 2-way pipeline parallelism using Llama 3.1 8B
-uv run torchrun --nproc_per_node=2 examples/llm/finetune.py \
+uv run torchrun --nproc_per_node=2 examples/llm_finetune/finetune.py \
     --config examples/llm_finetune/llama3_1/llama3_1_8b_hellaswag_pp.yaml
 ```
 
@@ -579,8 +579,8 @@ You can easily add pipeline parallelism to any existing training configuration t
 Add pipeline parallelism to an existing config using command-line arguments:
 
 ```bash
-uv run torchrun --nproc_per_node=2 examples/llm/finetune.py \
-    --config examples/llm/llama_3_2_1b_squad.yaml \
+uv run torchrun --nproc_per_node=2 examples/llm_finetune/finetune.py \
+    --config examples/llm_finetune/llama3_2/llama3_2_1b_squad.yaml \
     --distributed._target_ nemo_automodel.components.distributed.fsdp2.FSDP2Manager \
     --distributed.pp_size 2 \
     --autopipeline._target_ nemo_automodel.components.distributed.pipelining.AutoPipeline \
@@ -625,7 +625,7 @@ autopipeline:
 
 #### Pipeline + Data Parallelism (4 GPUs total)
 ```bash
-uv run torchrun --nproc_per_node=4 examples/llm/finetune.py \
+uv run torchrun --nproc_per_node=4 examples/llm_finetune/finetune.py \
     --config your_config.yaml \
     --distributed.pp_size 2 \
     --distributed.dp_size 2 \
@@ -634,7 +634,7 @@ uv run torchrun --nproc_per_node=4 examples/llm/finetune.py \
 
 #### Pipeline + Tensor Parallelism (4 GPUs total)
 ```bash
-uv run torchrun --nproc_per_node=4 examples/llm/finetune.py \
+uv run torchrun --nproc_per_node=4 examples/llm_finetune/finetune.py \
     --config your_config.yaml \
     --distributed.pp_size 2 \
     --distributed.tp_size 2 \
@@ -643,7 +643,7 @@ uv run torchrun --nproc_per_node=4 examples/llm/finetune.py \
 
 #### Full Hybrid: PP + DP + TP (8 GPUs total)
 ```bash
-uv run torchrun --nproc_per_node=8 examples/llm/finetune.py \
+uv run torchrun --nproc_per_node=8 examples/llm_finetune/finetune.py \
     --config your_config.yaml \
     --distributed.pp_size 2 \
     --distributed.dp_size 2 \
@@ -694,7 +694,7 @@ dataloader:
 Run training with:
 ```bash
 # Run with 2 GPUs for 2-way pipeline parallelism
-uv run torchrun --nproc_per_node=2 examples/llm/finetune.py --config config.yaml
+uv run torchrun --nproc_per_node=2 examples/llm_finetune/finetune.py --config config.yaml
 ```
 
 ## Troubleshooting
