mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-12-17 07:32:36 +00:00
Convert TGI
This commit is contained in:
Ashwin Bharambe
parent
9bb07ce298
commit
028530546f
14 changed files with 485 additions and 160 deletions
7
llama_stack/templates/tgi/__init__.py
Normal file
7
llama_stack/templates/tgi/__init__.py
Normal file
|
|
@ -0,0 +1,7 @@
|
|||
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
||||
# All rights reserved.
|
||||
#
|
||||
# This source code is licensed under the terms described in the LICENSE file in
|
||||
# the root directory of this source tree.
|
||||
|
||||
from .tgi import get_distribution_template # noqa: F401
|
||||
|
|
@ -1,12 +1,19 @@
|
|||
version: '2'
|
||||
name: tgi
|
||||
distribution_spec:
|
||||
description: Use TGI for running LLM inference
|
||||
description: Use (an external) TGI server for running LLM inference
|
||||
docker_image: llamastack/distribution-remote-tgi:test-0.0.52rc3
|
||||
providers:
|
||||
inference: remote::tgi
|
||||
inference:
|
||||
- remote::tgi
|
||||
memory:
|
||||
- inline::faiss
|
||||
- remote::chromadb
|
||||
- remote::pgvector
|
||||
safety: inline::llama-guard
|
||||
agents: inline::meta-reference
|
||||
telemetry: inline::meta-reference
|
||||
safety:
|
||||
- inline::llama-guard
|
||||
agents:
|
||||
- inline::meta-reference
|
||||
telemetry:
|
||||
- inline::meta-reference
|
||||
image_type: conda
|
||||
|
|
|
|||
125
llama_stack/templates/tgi/doc_template.md
Normal file
125
llama_stack/templates/tgi/doc_template.md
Normal file
|
|
@ -0,0 +1,125 @@
|
|||
# TGI Distribution
|
||||
|
||||
The `llamastack/distribution-{{ name }}` distribution consists of the following provider configurations.
|
||||
|
||||
{{ providers_table }}
|
||||
|
||||
You can use this distribution if you have GPUs and want to run an independent TGI server container for running inference.
|
||||
|
||||
{%- if docker_compose_env_vars %}
|
||||
### Environment Variables
|
||||
|
||||
The following environment variables can be configured:
|
||||
|
||||
{% for var, (default_value, description) in docker_compose_env_vars.items() %}
|
||||
- `{{ var }}`: {{ description }} (default: `{{ default_value }}`)
|
||||
{% endfor %}
|
||||
{% endif %}
|
||||
|
||||
{%- if default_models %}
|
||||
### Models
|
||||
|
||||
The following models are configured by default:
|
||||
{% for model in default_models %}
|
||||
- `{{ model.model_id }}`
|
||||
{% endfor %}
|
||||
{% endif %}
|
||||
|
||||
|
||||
## Using Docker Compose
|
||||
|
||||
You can use `docker compose` to start a TGI container and Llama Stack server container together.
|
||||
|
||||
```bash
|
||||
$ cd distributions/{{ name }}; 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://[::]:5001 (Press CTRL+C to quit)
|
||||
```
|
||||
|
||||
To kill the server
|
||||
```bash
|
||||
docker compose down
|
||||
```
|
||||
|
||||
|
||||
### Conda: TGI server + llama stack run
|
||||
|
||||
If you wish to separately spin up a TGI server, and connect with Llama Stack, you may use the following commands.
|
||||
|
||||
#### 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.
|
||||
|
||||
```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
|
||||
|
||||
**Via Conda**
|
||||
|
||||
```bash
|
||||
llama stack build --template {{ name }} --image-type conda
|
||||
# -- start a TGI server endpoint
|
||||
llama stack run ./gpu/run.yaml
|
||||
```
|
||||
|
||||
**Via Docker**
|
||||
```bash
|
||||
docker run --network host -it -p 5001:5001 \
|
||||
-v ./run.yaml:/root/my-run.yaml --gpus=all \
|
||||
llamastack/distribution-{{ name }} \
|
||||
--yaml_config /root/my-run.yaml
|
||||
```
|
||||
|
||||
We have provided a template `run.yaml` file in the `distributions/{{ name }}` 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
|
||||
config:
|
||||
url: http://127.0.0.1:5009
|
||||
```
|
||||
|
||||
|
||||
### (Optional) Update Model Serving Configuration
|
||||
To serve a new model with `tgi`, change the docker command flag `--model-id <model-to-serve>`.
|
||||
|
||||
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.
|
||||
|
||||
```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
|
||||
```bash
|
||||
docker run --rm -it -v $HOME/.cache/huggingface:/data \
|
||||
-p 5009:5009 --gpus all \
|
||||
ghcr.io/huggingface/text-generation-inference:latest \
|
||||
--dtype bfloat16 --usage-stats off --sharded false \
|
||||
--model-id meta-llama/Llama-3.2-3B-Instruct --port 5009
|
||||
```
|
||||
|
||||
In `run.yaml`, make sure you point the correct server endpoint to the TGI server endpoint serving your model.
|
||||
```yaml
|
||||
inference:
|
||||
- provider_id: tgi0
|
||||
provider_type: remote::tgi
|
||||
config:
|
||||
url: http://127.0.0.1:5009
|
||||
```
|
||||
97
llama_stack/templates/tgi/tgi.py
Normal file
97
llama_stack/templates/tgi/tgi.py
Normal file
|
|
@ -0,0 +1,97 @@
|
|||
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
||||
# All rights reserved.
|
||||
#
|
||||
# This source code is licensed under the terms described in the LICENSE file in
|
||||
# the root directory of this source tree.
|
||||
|
||||
from pathlib import Path
|
||||
|
||||
from llama_stack.distribution.datatypes import ModelInput, Provider, ShieldInput
|
||||
from llama_stack.providers.remote.inference.tgi import TGIImplConfig
|
||||
from llama_stack.templates.template import DistributionTemplate, RunConfigSettings
|
||||
|
||||
|
||||
def get_distribution_template() -> DistributionTemplate:
|
||||
providers = {
|
||||
"inference": ["remote::tgi"],
|
||||
"memory": ["inline::faiss", "remote::chromadb", "remote::pgvector"],
|
||||
"safety": ["inline::llama-guard"],
|
||||
"agents": ["inline::meta-reference"],
|
||||
"telemetry": ["inline::meta-reference"],
|
||||
}
|
||||
|
||||
inference_provider = Provider(
|
||||
provider_id="tgi-inference",
|
||||
provider_type="remote::tgi",
|
||||
config=TGIImplConfig.sample_run_config(
|
||||
url="${env.TGI_URL}",
|
||||
),
|
||||
)
|
||||
|
||||
inference_model = ModelInput(
|
||||
model_id="${env.INFERENCE_MODEL}",
|
||||
provider_id="tgi-inference",
|
||||
)
|
||||
safety_model = ModelInput(
|
||||
model_id="${env.SAFETY_MODEL}",
|
||||
provider_id="tgi-safety",
|
||||
)
|
||||
|
||||
return DistributionTemplate(
|
||||
name="tgi",
|
||||
distro_type="self_hosted",
|
||||
description="Use (an external) TGI server for running LLM inference",
|
||||
docker_image="llamastack/distribution-remote-tgi:test-0.0.52rc3",
|
||||
template_path=Path(__file__).parent / "doc_template.md",
|
||||
providers=providers,
|
||||
default_models=[inference_model, safety_model],
|
||||
run_configs={
|
||||
"run.yaml": RunConfigSettings(
|
||||
provider_overrides={
|
||||
"inference": [inference_provider],
|
||||
},
|
||||
default_models=[inference_model],
|
||||
),
|
||||
"run-with-safety.yaml": RunConfigSettings(
|
||||
provider_overrides={
|
||||
"inference": [
|
||||
inference_provider,
|
||||
Provider(
|
||||
provider_id="tgi-safety",
|
||||
provider_type="remote::tgi",
|
||||
config=TGIImplConfig.sample_run_config(
|
||||
url="${env.SAFETY_TGI_URL}",
|
||||
),
|
||||
),
|
||||
],
|
||||
},
|
||||
default_models=[
|
||||
inference_model,
|
||||
safety_model,
|
||||
],
|
||||
default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}")],
|
||||
),
|
||||
},
|
||||
docker_compose_env_vars={
|
||||
"LLAMASTACK_PORT": (
|
||||
"5001",
|
||||
"Port for the Llama Stack distribution server",
|
||||
),
|
||||
"INFERENCE_MODEL": (
|
||||
"meta-llama/Llama-3.2-3B-Instruct",
|
||||
"Inference model loaded into the TGI server",
|
||||
),
|
||||
"TGI_URL": (
|
||||
"http://host.docker.internal:8080}/v1",
|
||||
"URL of the TGI server with the main inference model",
|
||||
),
|
||||
"SAFETY_TGI_URL": (
|
||||
"http://host.docker.internal:8081/v1",
|
||||
"URL of the TGI server with the safety model",
|
||||
),
|
||||
"SAFETY_MODEL": (
|
||||
"meta-llama/Llama-Guard-3-1B",
|
||||
"Name of the safety (Llama-Guard) model to use",
|
||||
),
|
||||
},
|
||||
)
|
||||
Loading…
Add table
Add a link
Reference in a new issue