phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-17 15:02:37 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Convert TGI

Browse source
This commit is contained in:
Ashwin Bharambe 2024-11-17 14:49:41 -08:00
parent 9bb07ce298
commit 028530546f
14 changed files with 485 additions and 160 deletions
Download patch file Download diff file Expand all files Collapse all files

80
docs/source/getting_started/distributions/self_hosted_distro/tgi.md
View file

@ -2,35 +2,55 @@
The `llamastack/distribution-tgi` distribution consists of the following provider configurations.
| **API** | **Inference** | **Agents** | **Memory** | **Safety** | **Telemetry** |
|----------------- |--------------- |---------------- |-------------------------------------------------- |---------------- |---------------- |
| **Provider(s)** | remote::tgi | meta-reference | meta-reference, remote::pgvector, remote::chroma | meta-reference | meta-reference |
Provider Configuration
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ API ┃ Provider(s) ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ agents │ `inline::meta-reference` │
│ inference │ `remote::tgi` │
│ memory │ `inline::faiss`, `remote::chromadb`, `remote::pgvector` │
│ safety │ `inline::llama-guard` │
│ telemetry │ `inline::meta-reference` │
└───────────┴─────────────────────────────────────────────────────────┘
### Docker: Start the Distribution (Single Node GPU)
You can use this distribution if you have GPUs and want to run an independent TGI server container for running inference.### Environment Variables
> [!NOTE]
> This assumes you have access to GPU to start a TGI server with access to your GPU.
The following environment variables can be configured:
- `LLAMASTACK_PORT`: Port for the Llama Stack distribution server (default: `5001`)
- `INFERENCE_MODEL`: Inference model loaded into the TGI server (default: `meta-llama/Llama-3.2-3B-Instruct`)
- `TGI_URL`: URL of the TGI server with the main inference model (default: `http://host.docker.internal:8080}/v1`)
- `SAFETY_TGI_URL`: URL of the TGI server with the safety model (default: `http://host.docker.internal:8081/v1`)
- `SAFETY_MODEL`: Name of the safety (Llama-Guard) model to use (default: `meta-llama/Llama-Guard-3-1B`)
### Models
The following models are configured by default:
- `${env.INFERENCE_MODEL}`
- `${env.SAFETY_MODEL}`
```
$ cd distributions/tgi && docker compose up
## Using Docker Compose
You can use `docker compose` to start a TGI container and Llama Stack server container together.
```bash
$ cd distributions/tgi; docker compose up
```
The script will first start up TGI server, then start up Llama Stack distribution server hooking up to the remote TGI provider for inference. You should be able to see the following outputs --
```
```bash
[text-generation-inference] | 2024-10-15T18:56:33.810397Z INFO text_generation_router::server: router/src/server.rs:1813: Using config Some(Llama)
[text-generation-inference] | 2024-10-15T18:56:33.810448Z WARN text_generation_router::server: router/src/server.rs:1960: Invalid hostname, defaulting to 0.0.0.0
[text-generation-inference] | 2024-10-15T18:56:33.864143Z INFO text_generation_router::server: router/src/server.rs:2353: Connected
INFO: Started server process [1]
INFO: Waiting for application startup.
INFO: Application startup complete.
INFO: Uvicorn running on http://[::]:5000 (Press CTRL+C to quit)
INFO: Uvicorn running on http://[::]:5001 (Press CTRL+C to quit)
```
To kill the server
```
```bash
docker compose down
```
@ -42,8 +62,12 @@ If you wish to separately spin up a TGI server, and connect with Llama Stack, yo
#### Start TGI server locally
- Please check the [TGI Getting Started Guide](https://github.com/huggingface/text-generation-inference?tab=readme-ov-file#get-started) to get a TGI endpoint.
```
docker run --rm -it -v $HOME/.cache/huggingface:/data -p 5009:5009 --gpus all ghcr.io/huggingface/text-generation-inference:latest --dtype bfloat16 --usage-stats on --sharded false --model-id meta-llama/Llama-3.1-8B-Instruct --port 5009
```bash
docker run --rm -it -v $HOME/.cache/huggingface:/data \
-p 5009:5009 --gpus all \
ghcr.io/huggingface/text-generation-inference:latest \
--dtype bfloat16 --usage-stats on --sharded false \
--model-id meta-llama/Llama-3.2-3B-Instruct --port 5009
```
#### Start Llama Stack server pointing to TGI server
@ -57,12 +81,15 @@ llama stack run ./gpu/run.yaml
```
**Via Docker**
```
docker run --network host -it -p 5000:5000 -v ./run.yaml:/root/my-run.yaml --gpus=all llamastack/distribution-tgi --yaml_config /root/my-run.yaml
```bash
docker run --network host -it -p 5001:5001 \
-v ./run.yaml:/root/my-run.yaml --gpus=all \
llamastack/distribution-tgi \
--yaml_config /root/my-run.yaml
```
Make sure in you `run.yaml` file, you inference provider is pointing to the correct TGI server endpoint. E.g.
```
We have provided a template `run.yaml` file in the `distributions/tgi` directory. Make sure in your `run.yaml` file, you inference provider is pointing to the correct TGI server endpoint. E.g.
```yaml
inference:
- provider_id: tgi0
provider_type: remote::tgi
@ -76,17 +103,24 @@ To serve a new model with `tgi`, change the docker command flag `--model-id <mod
This can be done by edit the `command` args in `compose.yaml`. E.g. Replace "Llama-3.2-1B-Instruct" with the model you want to serve.
```
command: ["--dtype", "bfloat16", "--usage-stats", "on", "--sharded", "false", "--model-id", "meta-llama/Llama-3.2-1B-Instruct", "--port", "5009", "--cuda-memory-fraction", "0.3"]
```yaml
command: >
--dtype bfloat16 --usage-stats on --sharded false
--model-id meta-llama/Llama-3.2-1B-Instruct
--port 5009 --cuda-memory-fraction 0.7
```
or by changing the docker run command's `--model-id` flag
```
docker run --rm -it -v $HOME/.cache/huggingface:/data -p 5009:5009 --gpus all ghcr.io/huggingface/text-generation-inference:latest --dtype bfloat16 --usage-stats on --sharded false --model-id meta-llama/Llama-3.2-1B-Instruct --port 5009
```bash
docker run --rm -it -v $HOME/.cache/huggingface:/data \
-p 5009:5009 --gpus all \
ghcr.io/huggingface/text-generation-inference:latest \
--dtype bfloat16 --usage-stats off --sharded false \
--model-id meta-llama/Llama-3.2-3B-Instruct --port 5009
```
In `run.yaml`, make sure you point the correct server endpoint to the TGI server endpoint serving your model.
```
```yaml
inference:
- provider_id: tgi0
provider_type: remote::tgi