# What does this PR do?
[Provide a short summary of what this PR does and why. Link to relevant
issues if applicable.]
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
[//]: # (## Documentation)
---------
Co-authored-by: Francisco Arceo <farceo@redhat.com>
# What does this PR do?
This stubs in some OpenAI server-side compatibility with three new
endpoints:
/v1/openai/v1/models
/v1/openai/v1/completions
/v1/openai/v1/chat/completions
This gives common inference apps using OpenAI clients the ability to
talk to Llama Stack using an endpoint like
http://localhost:8321/v1/openai/v1 .
The two "v1" instances in there isn't awesome, but the thinking is that
Llama Stack's API is v1 and then our OpenAI compatibility layer is
compatible with OpenAI V1. And, some OpenAI clients implicitly assume
the URL ends with "v1", so this gives maximum compatibility.
The openai models endpoint is implemented in the routing layer, and just
returns all the models Llama Stack knows about.
The following providers should be working with the new OpenAI
completions and chat/completions API:
* remote::anthropic (untested)
* remote::cerebras-openai-compat (untested)
* remote::fireworks (tested)
* remote::fireworks-openai-compat (untested)
* remote::gemini (untested)
* remote::groq-openai-compat (untested)
* remote::nvidia (tested)
* remote::ollama (tested)
* remote::openai (untested)
* remote::passthrough (untested)
* remote::sambanova-openai-compat (untested)
* remote::together (tested)
* remote::together-openai-compat (untested)
* remote::vllm (tested)
The goal to support this for every inference provider - proxying
directly to the provider's OpenAI endpoint for OpenAI-compatible
providers. For providers that don't have an OpenAI-compatible API, we'll
add a mixin to translate incoming OpenAI requests to Llama Stack
inference requests and translate the Llama Stack inference responses to
OpenAI responses.
This is related to #1817 but is a bit larger in scope than just chat
completions, as I have real use-cases that need the older completions
API as well.
## Test Plan
### vLLM
```
VLLM_URL="http://localhost:8000/v1" INFERENCE_MODEL="meta-llama/Llama-3.2-3B-Instruct" llama stack build --template remote-vllm --image-type venv --run
LLAMA_STACK_CONFIG=http://localhost:8321 INFERENCE_MODEL="meta-llama/Llama-3.2-3B-Instruct" python -m pytest -v tests/integration/inference/test_openai_completion.py --text-model "meta-llama/Llama-3.2-3B-Instruct"
```
### ollama
```
INFERENCE_MODEL="llama3.2:3b-instruct-q8_0" llama stack build --template ollama --image-type venv --run
LLAMA_STACK_CONFIG=http://localhost:8321 INFERENCE_MODEL="llama3.2:3b-instruct-q8_0" python -m pytest -v tests/integration/inference/test_openai_completion.py --text-model "llama3.2:3b-instruct-q8_0"
```
## Documentation
Run a Llama Stack distribution that uses one of the providers mentioned
in the list above. Then, use your favorite OpenAI client to send
completion or chat completion requests with the base_url set to
http://localhost:8321/v1/openai/v1 . Replace "localhost:8321" with the
host and port of your Llama Stack server, if different.
---------
Signed-off-by: Ben Browning <bbrownin@redhat.com>
# What does this PR do?
Incorporating some feedback into the docs.
- **`docs/source/getting_started/index.md`:**
- Demo actually does RAG now
- Simplified the installation command for dependencies.
- Updated demo script examples to align with the latest API changes.
- Replaced manual document manipulation with `RAGDocument` for clarity
and maintainability.
- Introduced new logic for model and embedding selection using the Llama
Stack Client SDK.
- Enhanced examples to showcase proper agent initialization and logging.
- **`docs/source/getting_started/detailed_tutorial.md`:**
- Updated the section for listing models to include proper code
formatting with `bash`.
- Removed and reorganized the "Run the Demos" section for clarity.
- Adjusted tab-item structures and added new instructions for demo
scripts.
- **`docs/_static/css/my_theme.css`:**
- Updated heading styles to include `h2`, `h3`, and `h4` for consistent
font weight.
- Added a new style for `pre` tags to wrap text and break long words,
this is particularly useful for rendering long output from generation.
## Test Plan
Tested locally. Screenshot for reference:
<img width="1250" alt="Screenshot 2025-04-10 at 10 12 12 PM"
src="https://github.com/user-attachments/assets/ce1c8986-e072-4c6f-a697-ed0d8fb75b34"
/>
---------
Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>
# What does this PR do?
[Provide a short summary of what this PR does and why. Link to relevant
issues if applicable.]
Fixes a couple of errors in PVC/Secret setup and adds context for
expected Hugging Face token
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
[//]: # (## Documentation)
# What does this PR do?
Another doc enhancement for
https://github.com/meta-llama/llama-stack/issues/1818
Summary of changes:
- `docs/source/distributions/configuration.md`
- Updated dropdown title to include a more user-friendly description.
- `docs/_static/css/my_theme.css`
- Added styling for `<h3>` elements to set a normal font weight.
- `docs/source/distributions/starting_llama_stack_server.md`
- Changed section headers from bold text to proper markdown headers
(e.g., `##`).
- Improved descriptions for starting Llama Stack server using different
methods (library, container, conda, Kubernetes).
- Enhanced clarity and structure by converting instructions into
markdown headers and improved formatting.
- `docs/source/getting_started/index.md`
- Major restructuring of the "Quick Start" guide:
- Added new introductory section for Llama Stack and its capabilities.
- Reorganized steps into clearer subsections with proper markdown
headers.
- Replaced dropdowns with tabbed content for OS-specific instructions.
- Added detailed steps for setting up and running the Llama Stack server
and client.
- Introduced new sections for running basic inference and building
agents.
- Enhanced readability and visual structure with emojis, admonitions,
and examples.
- `docs/source/providers/index.md`
- Updated the list of LLM inference providers to include "Ollama."
- Expanded the list of vector databases to include "SQLite-Vec."
Let me know if you need further details!
## Test Plan
Renders locally, included screenshot.
# Documentation
For https://github.com/meta-llama/llama-stack/issues/1818
<img width="1332" alt="Screenshot 2025-04-09 at 11 07 12 AM"
src="https://github.com/user-attachments/assets/c106efb9-076c-4059-a4e0-a30fa738585b"
/>
---------
Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>
# What does this PR do?
* Manage UI deps in pyproject
* Use a new "ui" dep group to pull the deps with "uv"
* Simplify the run command
* Bump versions in requirements.txt
Signed-off-by: Sébastien Han <seb@redhat.com>
# What does this PR do?
1. Adding some lightweight JS to detect the default browser setting for
dark/light mode
3. Setting default screen setting to light mode as to not change default
behavior.
From the docs: https://github.com/MrDogeBro/sphinx_rtd_dark_mode
>This lets you choose which theme the user sees when they load the docs
for the first time ever. After the first time however, this setting has
no effect as the users preference is stored in local storage within
their browser. This option accepts a boolean for the value. If this
option is true (the default option), users will start in dark mode when
first visiting the site. If this option is false, users will start in
light mode when they first visit the site.
# Closes#1915
## Test Plan
Tested locally on my Mac on Safari and Chrome.
---------
Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>
# What does this PR do?
download the getting started w/ ollama model instead of downloading and
running it.
directly running it was necessary before
https://github.com/meta-llama/llama-stack/pull/1854
## Test Plan
run the code on the page
# What does this PR do?
Providers that live outside of the llama-stack codebase are now
supported.
A new property `external_providers_dir` has been added to the main
config and can be configured as follow:
```
external_providers_dir: /etc/llama-stack/providers.d/
```
Where the expected structure is:
```
providers.d/
inference/
custom_ollama.yaml
vllm.yaml
vector_io/
qdrant.yaml
```
Where `custom_ollama.yaml` is:
```
adapter:
adapter_type: custom_ollama
pip_packages: ["ollama", "aiohttp"]
config_class: llama_stack_ollama_provider.config.OllamaImplConfig
module: llama_stack_ollama_provider
api_dependencies: []
optional_api_dependencies: []
```
Obviously the package must be installed on the system, here is the
`llama_stack_ollama_provider` example:
```
$ uv pip show llama-stack-ollama-provider
Using Python 3.10.16 environment at: /Users/leseb/Documents/AI/llama-stack/.venv
Name: llama-stack-ollama-provider
Version: 0.1.0
Location: /Users/leseb/Documents/AI/llama-stack/.venv/lib/python3.10/site-packages
Editable project location: /private/var/folders/mq/rnm5w_7s2d3fxmtkx02knvhm0000gn/T/tmp.ZBHU5Ezxg4/ollama/llama-stack-ollama-provider
Requires:
Required-by:
```
Closes: https://github.com/meta-llama/llama-stack/issues/658
Signed-off-by: Sébastien Han <seb@redhat.com>
Add the content to use AMD GPU as the vLLM server. Split the original
part to two sub chapters,
1. AMD vLLM server
2. NVIDIA vLLM server (orignal)
# What does this PR do?
[Provide a short summary of what this PR does and why. Link to relevant
issues if applicable.]
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
[//]: # (## Documentation)
---------
Signed-off-by: Alex He <alehe@amd.com>
# What does this PR do?
## Test Plan
export MODEL=accounts/fireworks/models/llama4-scout-instruct-basic;
LLAMA_STACK_CONFIG=verification pytest -s -v tests/integration/inference
--vision-model $MODEL --text-model $MODEL
# What does this PR do?
[Provide a short summary of what this PR does and why. Link to relevant
issues if applicable.]
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
[//]: # (## Documentation)
# What does this PR do?
This PR modifies some of the docs to help them map to (1) the mental
model of software engineers building AI models starting with RAG and
then moving to Agents and (2) aligning the navbar somewhat closer to the
diagram on the home page.
## Test Plan
N/A Tested locally.
# Documentation
Take a look at the screen shot for below and after.
## Before
## After
---------
Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>
# What does this PR do?
Another for https://github.com/meta-llama/llama-stack/issues/1815
This links the `CONTRIBUTING.md` file directly so that we don't have to
maintain two different files.
Also I updated the title for RAG under Building AI Applications.
## Changes
Look of what the Contributing page looks like, proof it sources directly
from the markdown file.
---------
Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>
# What does this PR do?
Don't return list for runtime tools. Instead return Response object for
pagination and consistency with other APIs.
---------
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
# What does this PR do?
Move pagination logic from LocalFS and HuggingFace implementations into
a common helper function to ensure consistent pagination behavior across
providers. This reduces code duplication and centralizes pagination
logic in one place.
## Test Plan
Run this script:
```
from llama_stack_client import LlamaStackClient
# Initialize the client
client = LlamaStackClient(base_url="http://localhost:8321")
# Register a dataset
response = client.datasets.register(
purpose="eval/messages-answer", # or "eval/question-answer" or "post-training/messages"
source={"type": "uri", "uri": "huggingface://datasets/llamastack/simpleqa?split=train"},
dataset_id="my_dataset", # optional, will be auto-generated if not provided
metadata={"description": "My evaluation dataset"}, # optional
)
# Verify the dataset was registered by listing all datasets
datasets = client.datasets.list()
print(f"Registered datasets: {[d.identifier for d in datasets]}")
# You can then access the data using the datasetio API
# rows = client.datasets.iterrows(dataset_id="my_dataset", start_index=1, limit=2)
rows = client.datasets.iterrows(dataset_id="my_dataset")
print(f"Data: {rows.data}")
```
And play with `start_index` and `limit`.
[//]: # (## Documentation)
Signed-off-by: Sébastien Han <seb@redhat.com>
# What does this PR do?
The goal of this PR is to make the pages easier to navigate by surfacing
the child pages on the navbar, updating some of the copy, moving some of
the files around.
Some changes:
1. Clarifying Titles
2. Restructuring "Distributions" more formally in its own page to be
consistent with Providers and adding some clarity to the child pages to
surface them and make them easier to navigate
3. Updated sphinx config to not collapse navigation by default
4. Updated copyright year to be calculated dynamically
5. Moved `docs/source/distributions/index.md` ->
`docs/source/distributions/starting_llama_stack_server.md`
Another for https://github.com/meta-llama/llama-stack/issues/1815
## Test Plan
Tested locally and pages build (screen shots for example).
## Documentation
### Before:
### After:
Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>
# What does this PR do?
This PR updates the sink name configuration for traces and metrics in
LlamaStack to align with the latest changes introduced in version 0.1.8.
Previously, when using the `otel` sink along with other sinks (like
`console` and `sqlite`), the system threw a **ValueError**, with the
message:
```shell
Value error, 'otel' is not a valid TelemetrySink [type=value_error, input_value='console,otel,sqlite', input_type=str]
For further information visit https://errors.pydantic.dev/2.10/v/value_error
```
## Test Plan
- **Test 1:**
Ran the LlamaStack server with a configuration containing
`console,otel,sqlite` as sinks.
- **Expected result:** No errors related to invalid sink names.
- **Result:** The system ran without throwing a `ValueError`.
- **Test 2:**
Verified that the `otel_trace`, `otel_metric` sink now works in
combination with other sinks (`console`, `sqlite`).
- **Expected result:** Telemetry data is correctly sent to all specified
sinks without errors.
- **Result:** All telemetry data was successfully sent to the specified
sinks.
# What does this PR do?
This is the second attempt to switch to system packages by default. Now
with a hack to detect conda environment - in which case conda image-type
is used.
Note: Conda will only be used when --image-name is unset *and*
CONDA_DEFAULT_ENV is set. This means that users without conda will
correctly fall back to using system packages when no --image-* arguments
are passed at all.
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
Uses virtualenv:
```
$ llama stack build --template ollama --image-type venv
$ llama stack run --image-type venv ~/.llama/distributions/ollama/ollama-run.yaml
[...]
Using virtual environment: /home/ec2-user/src/llama-stack/schedule/.local
[...]
```
Uses system packages (virtualenv already initialized):
```
$ llama stack run ~/.llama/distributions/ollama/ollama-run.yaml
[...]
INFO 2025-03-27 20:46:22,882 llama_stack.cli.stack.run:142 server: No image type or image name provided. Assuming environment packages.
[...]
```
Attempt to run from environment packages without necessary packages
installed:
```
$ python -m venv barebones
$ . ./barebones/bin/activate
$ pip install -e . # to install llama command
$ llama stack run ~/.llama/distributions/ollama/ollama-run.yaml
[...]
ModuleNotFoundError: No module named 'fastapi'
```
^ failed as expected because the environment doesn't have necessary
packages installed.
Now install some packages in the new environment:
```
$ pip install fastapi opentelemetry-api opentelemetry-sdk opentelemetry-exporter-otlp aiosqlite ollama openai datasets faiss-cpu mcp autoevals
$ llama stack run ~/.llama/distributions/ollama/ollama-run.yaml
[...]
Uvicorn running on http://['::', '0.0.0.0']:8321 (Press CTRL+C to quit)
```
Now see if setting CONDA_DEFAULT_ENV will change what happens by
default:
```
$ export CONDA_DEFAULT_ENV=base
$ llama stack run ~/.llama/distributions/ollama/ollama-run.yaml
[...]
Using conda environment: base
Conda environment base does not exist.
[...]
```
---------
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
# What does this PR do?
- hide distro doc (docker needs to be thoroughly tested).
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
- docs
[//]: # (## Documentation)
# What does this PR do?
No violators are currently in-tree. This is just hardening the api specs
for future consistency.
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
# What does this PR do?
* Fix location of `run.yaml` relative to the cloned llama stack
repository
* Drop `-it` from `docker run` commands as its not needed running
services
## Test Plan
* Verified running the llama stack following updated instruction
CC: @ashwinb
Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Fixes multiple issues
1. llama stack build of dependencies was breaking with incompatible
numpy / pandas when importing datasets
Moved the notebook to start a local server instead of using library as a
client. This way the setup is cleaner since its all contained and by
using `uv run --with` we can test both the server setup process too in
CI and release time.
2. The change to [1] surfaced some other issues
- running `llama stack run` was defaulting to conda env name
- provider data was not being managed properly
- Some notebook cells (telemetry for evals) were not updated with latest
changes
Fixed all the issues and update the notebook.
### Test
1. Manually run it all in local env
2. `pytest -v -s --nbval-lax docs/getting_started.ipynb`
# What does this PR do?
This is to stay consistent with other APIs.
This change registers files in API, even though there are still no
providers. Removing tests that require a provider existing for a merged
API to enable it in API layer.
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
[//]: # (## Documentation)
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
# What does this PR do?
- as title
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
notebook
[//]: # (## Documentation)
# What does this PR do?
This PR adds support for NVIDIA's NeMo Customizer API to the Llama Stack
post-training module. The integration enables users to fine-tune models
using NVIDIA's cloud-based customization service through a consistent
Llama Stack interface.
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
Yet to be done
Things pending under this PR:
- [x] Integration of fine-tuned model(new checkpoint) for inference with
nvidia llm distribution
- [x] distribution integration of API
- [x] Add test cases for customizer(In Progress)
- [x] Documentation
```
LLAMA_STACK_BASE_URL=http://localhost:5002 pytest -v tests/client-sdk/post_training/test_supervised_fine_tuning.py
============================================================================================================================================================================ test session starts =============================================================================================================================================================================
platform linux -- Python 3.10.0, pytest-8.3.4, pluggy-1.5.0 -- /home/ubuntu/llama-stack/.venv/bin/python
cachedir: .pytest_cache
metadata: {'Python': '3.10.0', 'Platform': 'Linux-6.8.0-1021-gcp-x86_64-with-glibc2.35', 'Packages': {'pytest': '8.3.4', 'pluggy': '1.5.0'}, 'Plugins': {'nbval': '0.11.0', 'metadata': '3.1.1', 'anyio': '4.8.0', 'html': '4.1.1', 'asyncio': '0.25.3'}}
rootdir: /home/ubuntu/llama-stack
configfile: pyproject.toml
plugins: nbval-0.11.0, metadata-3.1.1, anyio-4.8.0, html-4.1.1, asyncio-0.25.3
asyncio: mode=strict, asyncio_default_fixture_loop_scope=None
collected 2 items
tests/client-sdk/post_training/test_supervised_fine_tuning.py::test_post_training_provider_registration[txt=8B] PASSED [ 50%]
tests/client-sdk/post_training/test_supervised_fine_tuning.py::test_list_training_jobs[txt=8B] PASSED [100%]
======================================================================================================================================================================== 2 passed, 1 warning in 0.10s ========================================================================================================================================================================
```
cc: @mattf @dglogo @sumitb
---------
Co-authored-by: Ubuntu <ubuntu@llama-stack-customizer-dev-inst-2tx95fyisatvlic4we8hidx5tfj.us-central1-a.c.brevdevprod.internal>
# What does this PR do?
* Removes the use of `huggingface-cli`
* Simplifies HF cache mount path
* Simplifies vLLM server startup command
* Separates PVC/secret creation from deployment/service
* Fixes a typo: "pod" should be "deployment"
Signed-off-by: Yuan Tang <terrytangyuan@gmail.com>
# What does this PR do?
**What**
Instead of adhoc creating a vectordb and chunking when documents ae sent
as an attachment to agent turn, we directly pass raw text from document
into messages to model for user context, and let model perform
summarization directly.
This removes the magic behaviour, and yields better performance than
existing approach.
**Improved Performance**
- RAG lifecycle notebook
- Model: 0.3 factuality score
- (+ websearch) Agent: 0.44 factuality score
- (+ vector db) Agent: 0.3 factuality score
- (+ raw context) Agent: 0.6 factuality score
Closes https://github.com/meta-llama/llama-stack/issues/1478
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
- [NEW] added section in RAG lifecycle notebook shows better performance
<img width="840" alt="image"
src="https://github.com/user-attachments/assets/a0c4e816-809a-41c0-9124-89825983e3f5"
/>
[//]: # (## Documentation)
# What does this PR do?
- We cannot directly return a literal type
> Note: this is not final jobs API change
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
<img width="837" alt="image"
src="https://github.com/user-attachments/assets/18a17561-35f9-443d-987d-54afdd6ff40c"
/>
[//]: # (## Documentation)
# What does this PR do?
[Provide a short summary of what this PR does and why. Link to relevant
issues if applicable.]
Typo fix for `output_dir` flag and misspelling of aggregate
[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])
## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
N/A
[//]: # (## Documentation)
