Update all links (#254)

Signed-off-by: Igor Gitman <[email protected]>

README.md

@@ -4,17 +4,17 @@ NeMo-Skills is a collection of pipelines to improve "skills" of large language m
 We mainly focus on the ability to solve mathematical problems, but you can use our pipelines for many other tasks as well.
 Here are some of the things we support.
 
-- [Flexible inference](https://kipok.github.io/NeMo-Skills/basics/inference): Seamlessly switch between API providers, local server and large-scale slurm jobs for LLM inference.
-- [Multiple formats](https://kipok.github.io/NeMo-Skills/pipelines/checkpoint-conversion): Use any of the [NeMo](https://github.com/NVIDIA/NeMo), [vLLM](https://github.com/vllm-project/vllm)
+- [Flexible inference](https://nvidia.github.io/NeMo-Skills/basics/inference): Seamlessly switch between API providers, local server and large-scale slurm jobs for LLM inference.
+- [Multiple formats](https://nvidia.github.io/NeMo-Skills/pipelines/checkpoint-conversion): Use any of the [NeMo](https://github.com/NVIDIA/NeMo), [vLLM](https://github.com/vllm-project/vllm)
   and [TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM) servers and easily convert checkpoints from one format to another.
-- [Model evaluation](https://kipok.github.io/NeMo-Skills/pipelines/evaluation): Evaluate your models on many popular benchmarks
+- [Model evaluation](https://nvidia.github.io/NeMo-Skills/pipelines/evaluation): Evaluate your models on many popular benchmarks
     - Math problem solving: gsm8k, math, amc23, aime24, omni-math (and many more)
     - Coding skills: human-eval, mbpp
     - Chat/instruction following: ifeval, arena-hard
     - General knowledge: mmlu (generative)
-- [Model training](https://kipok.github.io/NeMo-Skills/pipelines/training): Train models at speed-of-light using [NeMo-Aligner](https://github.com/NVIDIA/NeMo-Aligner/).
+- [Model training](https://nvidia.github.io/NeMo-Skills/pipelines/training): Train models at speed-of-light using [NeMo-Aligner](https://github.com/NVIDIA/NeMo-Aligner/).
 
-You can find the full documentation [here](https://kipok.github.io/NeMo-Skills/).
+You can find the full documentation [here](https://nvidia.github.io/NeMo-Skills/).
 
 ## OpenMathInstruct-2
 
@@ -83,7 +83,7 @@ The models trained on this dataset achieve strong results on common mathematical
   </tr>
 </table>
 
-We provide all instructions to [fully reproduce our results](https://kipok.github.io/NeMo-Skills/openmathinstruct2).
+We provide all instructions to [fully reproduce our results](https://nvidia.github.io/NeMo-Skills/openmathinstruct2).
 
 See our [paper](https://arxiv.org/abs/2410.01560) for ablations studies and more details!
 

docs/basics/inference.md

@@ -46,10 +46,10 @@ Click on :material-plus-circle: symbols in the snippet below to learn more detai
     print(outputs[0]["generation"]) # (3)!
     ```
 
-    1.   Here we use [generic/default](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/default.yaml) config
-         and [llama3-instruct](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/template/llama3-instruct.yaml) template.
+    1.   Here we use [generic/default](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/default.yaml) config
+         and [llama3-instruct](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/template/llama3-instruct.yaml) template.
 
-         See [nemo_skills/prompt](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt) for more config/template options
+         See [nemo_skills/prompt](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt) for more config/template options
          or [create your own prompts](prompt-format.md)
 
 
@@ -96,10 +96,10 @@ Click on :material-plus-circle: symbols in the snippet below to learn more detai
 
          To use OpenAI models, use `OPENAI_API_KEY` and set `base_url=https://api.openai.com/v1`.
 
-    2.   Here we use [generic/default](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/default.yaml) config.
+    2.   Here we use [generic/default](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/default.yaml) config.
          Note that with API models we can't add special tokens, so prompt template is not specified.
 
-         See [nemo_skills/prompt](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt) for more config/template options
+         See [nemo_skills/prompt](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt) for more config/template options
          or [create your own prompts](prompt-format.md)
 
 
@@ -138,12 +138,12 @@ Click on :material-plus-circle: symbols in the snippet below to learn more detai
     print(outputs[0]["generation"]) # (5)!
     ```
 
-    1.   Here we use [generic/default](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/default.yaml) config
-         and [llama3-instruct](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/template/llama3-instruct.yaml) template.
+    1.   Here we use [generic/default](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/default.yaml) config
+         and [llama3-instruct](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/template/llama3-instruct.yaml) template.
 
          Note how we are updating system message on the next line (you can also include it in the config directly).
 
-         See [nemo_skills/prompt](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt) for more config/template options
+         See [nemo_skills/prompt](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt) for more config/template options
          or [create your own prompts](prompt-format.md)
 
     2.   8B model doesn't always follow these instructions, so using 70B or 405B for code execution is recommended.

docs/basics/prerequisites.md

@@ -5,7 +5,7 @@
 To get started first install the repo (python 3.10+). Either clone and run `pip install -e .` or install directly with
 
 ```bash
-pip install git+https://github.com/Kipok/NeMo-Skills.git
+pip install git+https://github.com/NVIDIA/NeMo-Skills.git
 ```
 
 ## Environment variables
@@ -48,9 +48,9 @@ python -c "import nemo_skills; print(nemo_skills.__path__)"
 
 ## Cluster configs
 
-All of the [pipeline scripts](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/pipeline) accept `--cluster` argument which you can use
+All of the [pipeline scripts](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/pipeline) accept `--cluster` argument which you can use
 to control where the job gets executed. That argument picks up one of the configs inside your local
-[cluster_configs](https://github.com/Kipok/NeMo-Skills/tree/main/cluster_configs)
+[cluster_configs](https://github.com/NVIDIA/NeMo-Skills/tree/main/cluster_configs)
 folder by default, but you can specify another location with `--config_dir` or set it in `NEMO_SKILLS_CONFIG_DIR` env variable.
 You can also use `NEMO_SKILLS_CONFIG` env variable instead of the `--cluster` parameter.
 The cluster config defines an executor (local or slurm), mounts for data/model access and (slurm-only) various parameters
@@ -78,7 +78,7 @@ all other parameters are directly passed into the underlying `nemo_skills/...` s
 
 ## Running pipelines
 
-All of the [pipeline scripts](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/pipeline) can be called in 3 equivalent ways.
+All of the [pipeline scripts](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/pipeline) can be called in 3 equivalent ways.
 As an example let's see how to run [evaluation](../pipelines/evaluation.md) on 10 samples from gsm8k and math benchmarks
 
 ```bash title="ns command-line entrypoint"
@@ -132,7 +132,7 @@ set up on your machine.
 
 All of our scripts assume that data or models are mounted inside the appropriate container so before running any
 commands make sure to modify
-[cluster_configs/example-local.yaml](https://github.com/Kipok/NeMo-Skills/tree/main/cluster_configs/example-local.yaml).
+[cluster_configs/example-local.yaml](https://github.com/NVIDIA/NeMo-Skills/tree/main/cluster_configs/example-local.yaml).
 It's convenient to rename it to local.yaml (so you can use `--cluster local`) after you defined necessary mounts.
 
 Most of our containers are quite heavy, so the first time you run a job that requires a large container, it will take
@@ -144,5 +144,5 @@ to cache them.
 If you're running on slurm, you need to define some additional information inside cluster config.
 
 Populate the commented out fields inside
-[cluster_configs/example-slurm.yaml](https://github.com/Kipok/NeMo-Skills/tree/main/cluster_configs/example-slurm.yaml).
+[cluster_configs/example-slurm.yaml](https://github.com/NVIDIA/NeMo-Skills/tree/main/cluster_configs/example-slurm.yaml).
 It's convenient to rename it to slurm.yaml (so you can use `--cluster slurm`) or a cluster name if you use multiple slurm clusters.

docs/basics/prompt-format.md

@@ -13,9 +13,9 @@ Our prompts are configured via two input yaml files: prompt template and prompt
 The template file defines model-specific special tokens, e.g. bos, turn tokens,
 user/assistant/system message, special tokens for code execution, etc. All of the
 templates that we support by default are available in
-[nemo_skills/prompt/template](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/template)
+[nemo_skills/prompt/template](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/template)
 folder. Here is an example template for
-[llama3-instruct](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/template/llama3-instruct.yaml) models:
+[llama3-instruct](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/template/llama3-instruct.yaml) models:
 
 ```yaml
 # Prompt specification for the original Llama3-instruct model
@@ -64,9 +64,9 @@ in the included folder. If you provide a full path, we will take the file you sp
 The prompt config contains user and system messages with placeholders for keys from a data file.
 The configs are model independent (any model can be used with any config).
 All of the configs that we support by default are available in
-[nemo_skills/prompt/config](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/config)
+[nemo_skills/prompt/config](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/config)
 folder. Here is an example prompt for
-[math evaluations](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/math.yaml):
+[math evaluations](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/math.yaml):
 
 ```yaml
 # default prompt for all math benchmarks (e.g. gsm8k, math)
@@ -91,15 +91,15 @@ user: |-
 Note that we use `{problem}`, `{solution}` and `{examples}` format strings here. The `{examples}` is a special
 key that will be used to include few shot examples you specify above (it's empty unless you add `++examples_type` or
 specify it in the config like e.g. in
-[llama3-gsm8k prompt](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/gsm8k.yaml)).
+[llama3-gsm8k prompt](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/config/generic/gsm8k.yaml)).
 All other keys will need to be specified when you call `prompt.fill`
 (more on that in the [prompt-api section](#prompt-api)) so that we can replace placeholders with actual input.
 
 The input for few shot examples always comes from one of the available example types in
-[here](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/few_shot_examples/__init__.py). E.g. in the
-[llama3-gnstruct/gsm8k](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/config/llama3-instruct/gsm8k.yaml)
+[here](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/few_shot_examples/__init__.py). E.g. in the
+[llama3-gnstruct/gsm8k](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/config/llama3-instruct/gsm8k.yaml)
 prompt the `gsm8k_standard_few_shot` examples from
-[here](https://github.com/Kipok/NeMo-Skills/tree/main/nemo_skills/prompt/few_shot_examples/gsm8k.py) are used.
+[here](https://github.com/NVIDIA/NeMo-Skills/tree/main/nemo_skills/prompt/few_shot_examples/gsm8k.py) are used.
 
 
 ## Prompt API
@@ -157,4 +157,4 @@ which outputs
 ]
 ```
 
-You can also have a look at the [tests](https://github.com/Kipok/NeMo-Skills/tests/test_prompts.py) to see more examples of using our prompt API.
+You can also have a look at the [tests](https://github.com/NVIDIA/NeMo-Skills/tests/test_prompts.py) to see more examples of using our prompt API.

docs/code-execution/sandbox.md

@@ -7,7 +7,7 @@ we provide a basic sandbox that we use to execute code and validate the correctn
 ## Local sandbox
 
 The default sandbox option used in our pipeline is a local docker container.
-Check out [nemo_skills/code_execution/local_sandbox](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/code_execution/local_sandbox)
+Check out [nemo_skills/code_execution/local_sandbox](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/code_execution/local_sandbox)
 for implementation details.
 
 Please note that our provided sandbox is not fully secure and you are strongly encouraged to

docs/index.md

@@ -4,7 +4,7 @@ hide:
   - toc
 ---
 
-[NeMo-Skills](https://github.com/Kipok/NeMo-Skills) is a collection of pipelines to improve "skills" of large language models.
+[NeMo-Skills](https://github.com/NVIDIA/NeMo-Skills) is a collection of pipelines to improve "skills" of large language models.
 We mainly focus on the ability to solve mathematical problems, but you can use our pipelines for many other tasks as well.
 Here are some of the things we support.
 

docs/openmathinstruct2/dataset.md

@@ -254,7 +254,7 @@ update_output_files("<path to workspace>/new-problems-solution-augmentation/", c
 
 Now all the data is generated and you can follow up by converting it to the SFT format.
 We remove the problems marked as contaminated.
-We also remove problems and solutions with length > 1024 Llama tokens. 
+We also remove problems and solutions with length > 1024 Llama tokens.
 To avoid the models from generating extremely short solutions, we remove solutions shorter than 200 characters.
 
 ```bash
@@ -276,4 +276,4 @@ python -m nemo_skills.training.prepare_sft_data \
 
 ## Dataset contamination explorer
 
-To reproduce our dataset contamination explorer demo refer to [dataset_explorer_demo/README.md](https://github.com/Kipok/NeMo-Skills/blob/main/dataset_explorer_demo/README.md)
+To reproduce our dataset contamination explorer demo refer to [dataset_explorer_demo/README.md](https://github.com/NVIDIA/NeMo-Skills/blob/main/dataset_explorer_demo/README.md)

docs/openmathinstruct2/index.md

@@ -13,6 +13,6 @@ for some of the commands to complete!
 !!! note
 
     If you want to reproduce results for [OpenMathInstruct-1: A 1.8 Million Math Instruction Tuning Dataset](https://arxiv.org/abs/2402.10176)
-    please check out [v0.1.1](https://github.com/Kipok/NeMo-Skills/blob/v0.1.1/docs/reproducing-results.md)
-    branch of the [NeMo-Skills](https://github.com/Kipok/NeMo-Skills) repo and read the instructions in there.
+    please check out [v0.1.1](https://github.com/NVIDIA/NeMo-Skills/blob/v0.1.1/docs/reproducing-results.md)
+    branch of the [NeMo-Skills](https://github.com/NVIDIA/NeMo-Skills) repo and read the instructions in there.
 

docs/pipelines/checkpoint-conversion.md

@@ -2,21 +2,21 @@
 
 !!! info
 
-    This pipeline starting script is [nemo_skills/pipeline/convert.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/pipeline/convert.py)
+    This pipeline starting script is [nemo_skills/pipeline/convert.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/pipeline/convert.py)
 
     All extra parameters are passed to one of the following scripts
 
     * For conversion to NeMo:
-        - If `--model_type=llama`: [nemo_skills/conversion/hf_to_nemo_llama.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_nemo_llama.py)
-        - If `--model_type=qwen`: [nemo_skills/conversion/hf_to_nemo_qwen.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_nemo_qwen.py)
+        - If `--model_type=llama`: [nemo_skills/conversion/hf_to_nemo_llama.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_nemo_llama.py)
+        - If `--model_type=qwen`: [nemo_skills/conversion/hf_to_nemo_qwen.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_nemo_qwen.py)
 
     * For conversion to TensorRT-LLM:
-        - If `--model_type=llama`: [nemo_skills/conversion/hf_to_trtllm_llama.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_trtllm_llama.py)
-        - If `--model_type=qwen`: [nemo_skills/conversion/hf_to_trtllm_qwen.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_trtllm_qwen.py)
+        - If `--model_type=llama`: [nemo_skills/conversion/hf_to_trtllm_llama.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_trtllm_llama.py)
+        - If `--model_type=qwen`: [nemo_skills/conversion/hf_to_trtllm_qwen.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/conversion/hf_to_trtllm_qwen.py)
 
     * For conversion to HuggingFace:
-        - If `--model_type=llama`: [nemo_skills/conversion/nemo_to_hf_llama.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/conversion/nemo_to_hf_llama.py)
-        - If `--model_type=qwen`: [nemo_skills/conversion/nemo_to_hf_qwen.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/conversion/nemo_to_hf_qwen.py)
+        - If `--model_type=llama`: [nemo_skills/conversion/nemo_to_hf_llama.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/conversion/nemo_to_hf_llama.py)
+        - If `--model_type=qwen`: [nemo_skills/conversion/nemo_to_hf_qwen.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/conversion/nemo_to_hf_qwen.py)
 
 
 We support 3 common model formats. Here are some recommendations on when each format should be used.

docs/pipelines/decontamination.md

@@ -2,9 +2,9 @@
 
 !!! info
 
-    This pipeline starting script is [nemo_skills/pipeline/check_contamination.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/pipeline/check_contamination.py)
+    This pipeline starting script is [nemo_skills/pipeline/check_contamination.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/pipeline/check_contamination.py)
 
-    All extra parameters are passed to [nemo_skills/inference/check_contamination.py](https://github.com/Kipok/NeMo-Skills/blob/main/nemo_skills/inference/check_contamination.py)
+    All extra parameters are passed to [nemo_skills/inference/check_contamination.py](https://github.com/NVIDIA/NeMo-Skills/blob/main/nemo_skills/inference/check_contamination.py)
 
 We implemented an LLM-based data decontamination pipeline following
 [lmsys methodology](https://lmsys.org/blog/2023-11-14-llm-decontaminator/).