mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-08-12 04:50:39 +00:00
refactor: remove Conda support from Llama Stack (#2969)
# What does this PR do? <!-- Provide a short summary of what this PR does and why. Link to relevant issues if applicable. --> This PR is responsible for removal of Conda support in Llama Stack <!-- If resolving an issue, uncomment and update the line below --> <!-- Closes #[issue-number] --> Closes #2539 ## 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.* -->
This commit is contained in:
parent
f2eee4e417
commit
a749d5f4a4
44 changed files with 159 additions and 311 deletions
|
@ -69,9 +69,6 @@ def run_stack_build_command(args: argparse.Namespace) -> None:
|
|||
if args.image_type == ImageType.VENV.value:
|
||||
current_venv = os.environ.get("VIRTUAL_ENV")
|
||||
image_name = args.image_name or current_venv
|
||||
elif args.image_type == ImageType.CONDA.value:
|
||||
current_conda_env = os.environ.get("CONDA_DEFAULT_ENV")
|
||||
image_name = args.image_name or current_conda_env
|
||||
else:
|
||||
image_name = args.image_name
|
||||
|
||||
|
@ -132,7 +129,7 @@ def run_stack_build_command(args: argparse.Namespace) -> None:
|
|||
)
|
||||
if not args.image_type:
|
||||
cprint(
|
||||
f"Please specify a image-type (container | conda | venv) for {args.template}",
|
||||
f"Please specify a image-type (container | venv) for {args.template}",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
|
@ -158,22 +155,7 @@ def run_stack_build_command(args: argparse.Namespace) -> None:
|
|||
),
|
||||
)
|
||||
|
||||
if image_type == ImageType.CONDA.value:
|
||||
if not image_name:
|
||||
cprint(
|
||||
f"No current conda environment detected or specified, will create a new conda environment with the name `llamastack-{name}`",
|
||||
color="yellow",
|
||||
file=sys.stderr,
|
||||
)
|
||||
image_name = f"llamastack-{name}"
|
||||
else:
|
||||
cprint(
|
||||
f"Using conda environment {image_name}",
|
||||
color="green",
|
||||
file=sys.stderr,
|
||||
)
|
||||
else:
|
||||
image_name = f"llamastack-{name}"
|
||||
image_name = f"llamastack-{name}"
|
||||
|
||||
cprint(
|
||||
textwrap.dedent(
|
||||
|
@ -372,10 +354,7 @@ def _run_stack_build_command_from_build_config(
|
|||
else:
|
||||
if not image_name:
|
||||
raise ValueError("Please specify an image name when building a container image without a template")
|
||||
elif build_config.image_type == LlamaStackImageType.CONDA.value:
|
||||
if not image_name:
|
||||
raise ValueError("Please specify an image name when building a conda image")
|
||||
elif build_config.image_type == LlamaStackImageType.VENV.value:
|
||||
else:
|
||||
if not image_name and os.environ.get("UV_SYSTEM_PYTHON"):
|
||||
image_name = "__system__"
|
||||
if not image_name:
|
||||
|
@ -431,7 +410,6 @@ def _run_stack_build_command_from_build_config(
|
|||
|
||||
return_code = build_image(
|
||||
build_config,
|
||||
build_file_path,
|
||||
image_name,
|
||||
template_or_config=template_name or config_path or str(build_file_path),
|
||||
run_config=run_config_file.as_posix() if run_config_file else None,
|
||||
|
|
|
@ -56,7 +56,7 @@ class StackBuild(Subcommand):
|
|||
"--image-name",
|
||||
type=str,
|
||||
help=textwrap.dedent(
|
||||
f"""[for image-type={"|".join(e.value for e in ImageType)}] Name of the conda or virtual environment to use for
|
||||
f"""[for image-type={"|".join(e.value for e in ImageType)}] Name of the virtual environment to use for
|
||||
the build. If not specified, currently active environment will be used if found.
|
||||
"""
|
||||
),
|
||||
|
|
|
@ -47,7 +47,8 @@ class StackRun(Subcommand):
|
|||
self.parser.add_argument(
|
||||
"--image-name",
|
||||
type=str,
|
||||
help="Name of the image to run.",
|
||||
default=None,
|
||||
help="Name of the image to run. Defaults to the current environment",
|
||||
)
|
||||
self.parser.add_argument(
|
||||
"--env",
|
||||
|
@ -58,7 +59,7 @@ class StackRun(Subcommand):
|
|||
self.parser.add_argument(
|
||||
"--image-type",
|
||||
type=str,
|
||||
help="Image Type used during the build. This can be either conda or container or venv.",
|
||||
help="Image Type used during the build. This can be only venv.",
|
||||
choices=[e.value for e in ImageType if e.value != ImageType.CONTAINER.value],
|
||||
)
|
||||
self.parser.add_argument(
|
||||
|
@ -67,20 +68,38 @@ class StackRun(Subcommand):
|
|||
help="Start the UI server",
|
||||
)
|
||||
|
||||
# If neither image type nor image name is provided, but at the same time
|
||||
# the current environment has conda breadcrumbs, then assume what the user
|
||||
# wants to use conda mode and not the usual default mode (using
|
||||
# pre-installed system packages).
|
||||
#
|
||||
# Note: yes, this is hacky. It's implemented this way to keep the existing
|
||||
# conda users unaffected by the switch of the default behavior to using
|
||||
# system packages.
|
||||
def _get_image_type_and_name(self, args: argparse.Namespace) -> tuple[str, str]:
|
||||
conda_env = os.environ.get("CONDA_DEFAULT_ENV")
|
||||
if conda_env and args.image_name == conda_env:
|
||||
logger.warning(f"Conda detected. Using conda environment {conda_env} for the run.")
|
||||
return ImageType.CONDA.value, args.image_name
|
||||
return args.image_type, args.image_name
|
||||
def _resolve_config_and_template(self, args: argparse.Namespace) -> tuple[Path | None, str | None]:
|
||||
"""Resolve config file path and template name from args.config"""
|
||||
from llama_stack.distribution.utils.config_dirs import DISTRIBS_BASE_DIR
|
||||
|
||||
if not args.config:
|
||||
return None, None
|
||||
|
||||
config_file = Path(args.config)
|
||||
has_yaml_suffix = args.config.endswith(".yaml")
|
||||
template_name = None
|
||||
|
||||
if not config_file.exists() and not has_yaml_suffix:
|
||||
# check if this is a template
|
||||
config_file = Path(REPO_ROOT) / "llama_stack" / "templates" / args.config / "run.yaml"
|
||||
if config_file.exists():
|
||||
template_name = args.config
|
||||
|
||||
if not config_file.exists() and not has_yaml_suffix:
|
||||
# check if it's a build config saved to ~/.llama dir
|
||||
config_file = Path(DISTRIBS_BASE_DIR / f"llamastack-{args.config}" / f"{args.config}-run.yaml")
|
||||
|
||||
if not config_file.exists():
|
||||
self.parser.error(
|
||||
f"File {str(config_file)} does not exist.\n\nPlease run `llama stack build` to generate (and optionally edit) a run.yaml file"
|
||||
)
|
||||
|
||||
if not config_file.is_file():
|
||||
self.parser.error(
|
||||
f"Config file must be a valid file path, '{config_file}' is not a file: type={type(config_file)}"
|
||||
)
|
||||
|
||||
return config_file, template_name
|
||||
|
||||
def _run_stack_run_cmd(self, args: argparse.Namespace) -> None:
|
||||
import yaml
|
||||
|
@ -90,7 +109,7 @@ class StackRun(Subcommand):
|
|||
|
||||
if args.enable_ui:
|
||||
self._start_ui_development_server(args.port)
|
||||
image_type, image_name = self._get_image_type_and_name(args)
|
||||
image_type, image_name = args.image_type, args.image_name
|
||||
|
||||
if args.config:
|
||||
try:
|
||||
|
@ -103,8 +122,8 @@ class StackRun(Subcommand):
|
|||
config_file = None
|
||||
|
||||
# Check if config is required based on image type
|
||||
if (image_type in [ImageType.CONDA.value, ImageType.VENV.value]) and not config_file:
|
||||
self.parser.error("Config file is required for venv and conda environments")
|
||||
if image_type == ImageType.VENV.value and not config_file:
|
||||
self.parser.error("Config file is required for venv environment")
|
||||
|
||||
if config_file:
|
||||
logger.info(f"Using run configuration: {config_file}")
|
||||
|
|
|
@ -8,7 +8,6 @@ from enum import Enum
|
|||
|
||||
|
||||
class ImageType(Enum):
|
||||
CONDA = "conda"
|
||||
CONTAINER = "container"
|
||||
VENV = "venv"
|
||||
|
||||
|
|
|
@ -7,7 +7,6 @@
|
|||
import importlib.resources
|
||||
import logging
|
||||
import sys
|
||||
from pathlib import Path
|
||||
|
||||
from pydantic import BaseModel
|
||||
from termcolor import cprint
|
||||
|
@ -106,7 +105,6 @@ def print_pip_install_help(config: BuildConfig):
|
|||
|
||||
def build_image(
|
||||
build_config: BuildConfig,
|
||||
build_file_path: Path,
|
||||
image_name: str,
|
||||
template_or_config: str,
|
||||
run_config: str | None = None,
|
||||
|
@ -138,18 +136,7 @@ def build_image(
|
|||
# build arguments
|
||||
if run_config is not None:
|
||||
args.extend(["--run-config", run_config])
|
||||
elif build_config.image_type == LlamaStackImageType.CONDA.value:
|
||||
script = str(importlib.resources.files("llama_stack") / "core/build_conda_env.sh")
|
||||
args = [
|
||||
script,
|
||||
"--env-name",
|
||||
str(image_name),
|
||||
"--build-file-path",
|
||||
str(build_file_path),
|
||||
"--normal-deps",
|
||||
" ".join(normal_deps),
|
||||
]
|
||||
elif build_config.image_type == LlamaStackImageType.VENV.value:
|
||||
else:
|
||||
script = str(importlib.resources.files("llama_stack") / "core/build_venv.sh")
|
||||
args = [
|
||||
script,
|
||||
|
|
|
@ -6,9 +6,6 @@
|
|||
# This source code is licensed under the terms described in the LICENSE file in
|
||||
# the root directory of this source tree.
|
||||
|
||||
# TODO: combine this with build_conda_env.sh since it is almost identical
|
||||
# the only difference is that we don't do any conda-specific setup
|
||||
|
||||
LLAMA_STACK_DIR=${LLAMA_STACK_DIR:-}
|
||||
LLAMA_STACK_CLIENT_DIR=${LLAMA_STACK_CLIENT_DIR:-}
|
||||
TEST_PYPI_VERSION=${TEST_PYPI_VERSION:-}
|
||||
|
@ -95,6 +92,8 @@ if [ -n "$LLAMA_STACK_CLIENT_DIR" ]; then
|
|||
echo "Using llama-stack-client-dir=$LLAMA_STACK_CLIENT_DIR"
|
||||
fi
|
||||
|
||||
ENVNAME=""
|
||||
|
||||
# pre-run checks to make sure we can proceed with the installation
|
||||
pre_run_checks() {
|
||||
local env_name="$1"
|
||||
|
|
|
@ -7,12 +7,10 @@
|
|||
# the root directory of this source tree.
|
||||
|
||||
cleanup() {
|
||||
envname="$1"
|
||||
|
||||
set +x
|
||||
echo "Cleaning up..."
|
||||
conda deactivate
|
||||
conda env remove --name "$envname" -y
|
||||
# For venv environments, no special cleanup is needed
|
||||
# This function exists to avoid "function not found" errors
|
||||
local env_name="$1"
|
||||
echo "Cleanup called for environment: $env_name"
|
||||
}
|
||||
|
||||
handle_int() {
|
||||
|
@ -31,19 +29,7 @@ handle_exit() {
|
|||
fi
|
||||
}
|
||||
|
||||
setup_cleanup_handlers() {
|
||||
trap handle_int INT
|
||||
trap handle_exit EXIT
|
||||
|
||||
if is_command_available conda; then
|
||||
__conda_setup="$('conda' 'shell.bash' 'hook' 2>/dev/null)"
|
||||
eval "$__conda_setup"
|
||||
conda deactivate
|
||||
else
|
||||
echo "conda is not available"
|
||||
exit 1
|
||||
fi
|
||||
}
|
||||
|
||||
# check if a command is present
|
||||
is_command_available() {
|
||||
|
|
|
@ -432,8 +432,8 @@ class BuildConfig(BaseModel):
|
|||
|
||||
distribution_spec: DistributionSpec = Field(description="The distribution spec to build including API providers. ")
|
||||
image_type: str = Field(
|
||||
default="conda",
|
||||
description="Type of package to build (conda | container | venv)",
|
||||
default="venv",
|
||||
description="Type of package to build (container | venv)",
|
||||
)
|
||||
image_name: str | None = Field(
|
||||
default=None,
|
||||
|
|
|
@ -40,7 +40,6 @@ port="$1"
|
|||
shift
|
||||
|
||||
SCRIPT_DIR=$(dirname "$(readlink -f "$0")")
|
||||
source "$SCRIPT_DIR/common.sh"
|
||||
|
||||
# Initialize variables
|
||||
yaml_config=""
|
||||
|
@ -75,9 +74,9 @@ while [[ $# -gt 0 ]]; do
|
|||
esac
|
||||
done
|
||||
|
||||
# Check if yaml_config is required based on env_type
|
||||
if [[ "$env_type" == "venv" || "$env_type" == "conda" ]] && [ -z "$yaml_config" ]; then
|
||||
echo -e "${RED}Error: --config is required for venv and conda environments${NC}" >&2
|
||||
# Check if yaml_config is required
|
||||
if [[ "$env_type" == "venv" ]] && [ -z "$yaml_config" ]; then
|
||||
echo -e "${RED}Error: --config is required for venv environment${NC}" >&2
|
||||
exit 1
|
||||
fi
|
||||
|
||||
|
@ -101,19 +100,14 @@ case "$env_type" in
|
|||
source "$env_path_or_name/bin/activate"
|
||||
fi
|
||||
;;
|
||||
"conda")
|
||||
if ! is_command_available conda; then
|
||||
echo -e "${RED}Error: conda not found" >&2
|
||||
exit 1
|
||||
fi
|
||||
eval "$(conda shell.bash hook)"
|
||||
conda deactivate && conda activate "$env_path_or_name"
|
||||
PYTHON_BINARY="$CONDA_PREFIX/bin/python"
|
||||
;;
|
||||
*)
|
||||
# Handle unsupported env_types here
|
||||
echo -e "${RED}Error: Unsupported environment type '$env_type'. Only 'venv' is supported.${NC}" >&2
|
||||
exit 1
|
||||
;;
|
||||
esac
|
||||
|
||||
if [[ "$env_type" == "venv" || "$env_type" == "conda" ]]; then
|
||||
if [[ "$env_type" == "venv" ]]; then
|
||||
set -x
|
||||
|
||||
if [ -n "$yaml_config" ]; then
|
||||
|
|
|
@ -9,7 +9,7 @@
|
|||
1. Start up Llama Stack API server. More details [here](https://llama-stack.readthedocs.io/en/latest/getting_started/index.html).
|
||||
|
||||
```
|
||||
llama stack build --template together --image-type conda
|
||||
llama stack build --template together --image-type venv
|
||||
|
||||
llama stack run together
|
||||
```
|
||||
|
|
|
@ -15,69 +15,21 @@ from termcolor import cprint
|
|||
log = logging.getLogger(__name__)
|
||||
|
||||
import importlib
|
||||
import json
|
||||
from pathlib import Path
|
||||
|
||||
from llama_stack.core.utils.image_types import LlamaStackImageType
|
||||
|
||||
|
||||
def formulate_run_args(image_type: str, image_name: str) -> list[str]:
|
||||
env_name = ""
|
||||
def formulate_run_args(image_type: str, image_name: str) -> list:
|
||||
# Only venv is supported now
|
||||
current_venv = os.environ.get("VIRTUAL_ENV")
|
||||
env_name = image_name or current_venv
|
||||
if not env_name:
|
||||
cprint(
|
||||
"No current virtual environment detected, please specify a virtual environment name with --image-name",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
return []
|
||||
|
||||
if image_type == LlamaStackImageType.CONDA.value:
|
||||
current_conda_env = os.environ.get("CONDA_DEFAULT_ENV")
|
||||
env_name = image_name or current_conda_env
|
||||
if not env_name:
|
||||
cprint(
|
||||
"No current conda environment detected, please specify a conda environment name with --image-name",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
return
|
||||
|
||||
def get_conda_prefix(env_name):
|
||||
# Conda "base" environment does not end with "base" in the
|
||||
# prefix, so should be handled separately.
|
||||
if env_name == "base":
|
||||
return os.environ.get("CONDA_PREFIX")
|
||||
# Get conda environments info
|
||||
conda_env_info = json.loads(subprocess.check_output(["conda", "info", "--envs", "--json"]).decode())
|
||||
envs = conda_env_info["envs"]
|
||||
for envpath in envs:
|
||||
if os.path.basename(envpath) == env_name:
|
||||
return envpath
|
||||
return None
|
||||
|
||||
cprint(f"Using conda environment: {env_name}", color="green", file=sys.stderr)
|
||||
conda_prefix = get_conda_prefix(env_name)
|
||||
if not conda_prefix:
|
||||
cprint(
|
||||
f"Conda environment {env_name} does not exist.",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
return
|
||||
|
||||
build_file = Path(conda_prefix) / "llamastack-build.yaml"
|
||||
if not build_file.exists():
|
||||
cprint(
|
||||
f"Build file {build_file} does not exist.\n\nPlease run `llama stack build` or specify the correct conda environment name with --image-name",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
return
|
||||
else:
|
||||
# else must be venv since that is the only valid option left.
|
||||
current_venv = os.environ.get("VIRTUAL_ENV")
|
||||
env_name = image_name or current_venv
|
||||
if not env_name:
|
||||
cprint(
|
||||
"No current virtual environment detected, please specify a virtual environment name with --image-name",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
return
|
||||
cprint(f"Using virtual environment: {env_name}", file=sys.stderr)
|
||||
cprint(f"Using virtual environment: {env_name}", file=sys.stderr)
|
||||
|
||||
script = importlib.resources.files("llama_stack") / "core/start_stack.sh"
|
||||
run_args = [
|
||||
|
@ -93,7 +45,8 @@ def in_notebook():
|
|||
try:
|
||||
from IPython import get_ipython
|
||||
|
||||
if "IPKernelApp" not in get_ipython().config: # pragma: no cover
|
||||
ipython = get_ipython()
|
||||
if ipython is None or "IPKernelApp" not in ipython.config: # pragma: no cover
|
||||
return False
|
||||
except ImportError:
|
||||
return False
|
||||
|
|
|
@ -9,5 +9,4 @@ import enum
|
|||
|
||||
class LlamaStackImageType(enum.Enum):
|
||||
CONTAINER = "container"
|
||||
CONDA = "conda"
|
||||
VENV = "venv"
|
||||
|
|
|
@ -20,7 +20,7 @@ This provider enables dataset management using NVIDIA's NeMo Customizer service.
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type conda
|
||||
llama stack build --template nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
|
@ -18,7 +18,7 @@ This provider enables running inference using NVIDIA NIM.
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type conda
|
||||
llama stack build --template nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
|
@ -22,7 +22,7 @@ This provider enables fine-tuning of LLMs using NVIDIA's NeMo Customizer service
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type conda
|
||||
llama stack build --template nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
|
@ -19,7 +19,7 @@ This provider enables safety checks and guardrails for LLM interactions using NV
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type conda
|
||||
llama stack build --template nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
|
@ -47,8 +47,7 @@ distribution_spec:
|
|||
- provider_type: remote::tavily-search
|
||||
- provider_type: inline::rag-runtime
|
||||
- provider_type: remote::model-context-protocol
|
||||
image_type: conda
|
||||
image_name: ci-tests
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- aiosqlite
|
||||
- asyncpg
|
||||
|
|
|
@ -29,8 +29,7 @@ distribution_spec:
|
|||
- provider_type: remote::brave-search
|
||||
- provider_type: remote::tavily-search
|
||||
- provider_type: inline::rag-runtime
|
||||
image_type: conda
|
||||
image_name: dell
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- aiosqlite
|
||||
- sqlalchemy[asyncio]
|
||||
|
|
|
@ -28,8 +28,7 @@ distribution_spec:
|
|||
- provider_type: remote::tavily-search
|
||||
- provider_type: inline::rag-runtime
|
||||
- provider_type: remote::model-context-protocol
|
||||
image_type: conda
|
||||
image_name: meta-reference-gpu
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- aiosqlite
|
||||
- sqlalchemy[asyncio]
|
||||
|
|
|
@ -58,7 +58,7 @@ $ llama model list --downloaded
|
|||
|
||||
## Running the Distribution
|
||||
|
||||
You can do this via Conda (build code) or Docker which has a pre-built image.
|
||||
You can do this via venv or Docker which has a pre-built image.
|
||||
|
||||
### Via Docker
|
||||
|
||||
|
@ -92,12 +92,12 @@ docker run \
|
|||
--env SAFETY_MODEL=meta-llama/Llama-Guard-3-1B
|
||||
```
|
||||
|
||||
### Via Conda
|
||||
### Via venv
|
||||
|
||||
Make sure you have done `uv pip install llama-stack` and have the Llama Stack CLI available.
|
||||
|
||||
```bash
|
||||
llama stack build --template {{ name }} --image-type conda
|
||||
llama stack build --template {{ name }} --image-type venv
|
||||
llama stack run distributions/{{ name }}/run.yaml \
|
||||
--port 8321 \
|
||||
--env INFERENCE_MODEL=meta-llama/Llama-3.2-3B-Instruct
|
||||
|
|
|
@ -23,8 +23,7 @@ distribution_spec:
|
|||
- provider_type: inline::basic
|
||||
tool_runtime:
|
||||
- provider_type: inline::rag-runtime
|
||||
image_type: conda
|
||||
image_name: nvidia
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- aiosqlite
|
||||
- sqlalchemy[asyncio]
|
||||
|
|
|
@ -105,7 +105,7 @@ curl -X DELETE "$NEMO_URL/v1/deployment/model-deployments/meta/llama-3.1-8b-inst
|
|||
|
||||
## Running Llama Stack with NVIDIA
|
||||
|
||||
You can do this via Conda or venv (build code), or Docker which has a pre-built image.
|
||||
You can do this via venv (build code), or Docker which has a pre-built image.
|
||||
|
||||
### Via Docker
|
||||
|
||||
|
@ -124,17 +124,6 @@ docker run \
|
|||
--env NVIDIA_API_KEY=$NVIDIA_API_KEY
|
||||
```
|
||||
|
||||
### Via Conda
|
||||
|
||||
```bash
|
||||
INFERENCE_MODEL=meta-llama/Llama-3.1-8b-Instruct
|
||||
llama stack build --template nvidia --image-type conda
|
||||
llama stack run ./run.yaml \
|
||||
--port 8321 \
|
||||
--env NVIDIA_API_KEY=$NVIDIA_API_KEY \
|
||||
--env INFERENCE_MODEL=$INFERENCE_MODEL
|
||||
```
|
||||
|
||||
### Via venv
|
||||
|
||||
If you've set up your local development environment, you can also build the image using your local virtual environment.
|
||||
|
|
|
@ -32,8 +32,7 @@ distribution_spec:
|
|||
- provider_type: remote::tavily-search
|
||||
- provider_type: inline::rag-runtime
|
||||
- provider_type: remote::model-context-protocol
|
||||
image_type: conda
|
||||
image_name: open-benchmark
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- aiosqlite
|
||||
- sqlalchemy[asyncio]
|
||||
|
|
|
@ -18,8 +18,7 @@ distribution_spec:
|
|||
- provider_type: remote::tavily-search
|
||||
- provider_type: inline::rag-runtime
|
||||
- provider_type: remote::model-context-protocol
|
||||
image_type: conda
|
||||
image_name: postgres-demo
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- asyncpg
|
||||
- psycopg2-binary
|
||||
|
|
|
@ -47,8 +47,7 @@ distribution_spec:
|
|||
- provider_type: remote::tavily-search
|
||||
- provider_type: inline::rag-runtime
|
||||
- provider_type: remote::model-context-protocol
|
||||
image_type: conda
|
||||
image_name: starter
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- aiosqlite
|
||||
- asyncpg
|
||||
|
|
|
@ -29,6 +29,7 @@ from llama_stack.core.datatypes import (
|
|||
)
|
||||
from llama_stack.core.distribution import get_provider_registry
|
||||
from llama_stack.core.utils.dynamic import instantiate_class_type
|
||||
from llama_stack.core.utils.image_types import LlamaStackImageType
|
||||
from llama_stack.providers.utils.inference.model_registry import ProviderModelEntry
|
||||
from llama_stack.providers.utils.kvstore.config import SqliteKVStoreConfig
|
||||
from llama_stack.providers.utils.kvstore.config import get_pip_packages as get_kv_pip_packages
|
||||
|
@ -314,8 +315,7 @@ class DistributionTemplate(BaseModel):
|
|||
container_image=self.container_image,
|
||||
providers=build_providers,
|
||||
),
|
||||
image_type="conda",
|
||||
image_name=self.name,
|
||||
image_type=LlamaStackImageType.VENV.value, # default to venv
|
||||
additional_pip_packages=sorted(set(additional_pip_packages)),
|
||||
)
|
||||
|
||||
|
|
|
@ -35,16 +35,11 @@ distribution_spec:
|
|||
- provider_id: braintrust
|
||||
provider_type: inline::braintrust
|
||||
tool_runtime:
|
||||
- provider_id: brave-search
|
||||
provider_type: remote::brave-search
|
||||
- provider_id: tavily-search
|
||||
provider_type: remote::tavily-search
|
||||
- provider_id: rag-runtime
|
||||
provider_type: inline::rag-runtime
|
||||
- provider_id: model-context-protocol
|
||||
provider_type: remote::model-context-protocol
|
||||
image_type: conda
|
||||
image_name: watsonx
|
||||
- provider_type: remote::brave-search
|
||||
- provider_type: remote::tavily-search
|
||||
- provider_type: inline::rag-runtime
|
||||
- provider_type: remote::model-context-protocol
|
||||
image_type: venv
|
||||
additional_pip_packages:
|
||||
- sqlalchemy[asyncio]
|
||||
- aiosqlite
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue