Add centml as remote inference provider

llama_stack/templates/centml/__init__.py

@@ -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 .centml import get_distribution_template  # noqa: F401

llama_stack/templates/centml/build.yaml

@@ -0,0 +1,32 @@
+version: '2'
+name: centml
+distribution_spec:
+  description: Use CentML for running LLM inference
+  providers:
+    inference:
+    - remote::centml
+    memory:
+    - inline::faiss
+    - remote::chromadb
+    - remote::pgvector
+    safety:
+    - inline::llama-guard
+    agents:
+    - inline::meta-reference
+    telemetry:
+    - inline::meta-reference
+    eval:
+    - inline::meta-reference
+    datasetio:
+    - remote::huggingface
+    - inline::localfs
+    scoring:
+    - inline::basic
+    - inline::llm-as-judge
+    - inline::braintrust
+    tool_runtime:
+    - remote::brave-search
+    - remote::tavily-search
+    - inline::code-interpreter
+    - inline::memory-runtime
+image_type: conda

llama_stack/templates/centml/centml.py

@@ -0,0 +1,110 @@
+# 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_models.sku_list import all_registered_models
+from llama_stack.apis.models.models import ModelType
+from llama_stack.distribution.datatypes import ModelInput, Provider, ShieldInput
+from llama_stack.providers.inline.inference.sentence_transformers import (
+    SentenceTransformersInferenceConfig,
+)
+from llama_stack.providers.inline.memory.faiss.config import FaissImplConfig
+from llama_stack.providers.remote.inference.centml.config import CentMLImplConfig
+
+# If your CentML adapter has a MODEL_ALIASES constant with known model mappings:
+from llama_stack.providers.remote.inference.centml.centml import MODEL_ALIASES
+
+from llama_stack.templates.template import DistributionTemplate, RunConfigSettings
+
+
+def get_distribution_template() -> DistributionTemplate:
+    """
+    Returns a distribution template for running Llama Stack with CentML inference.
+    """
+    providers = {
+        "inference": ["remote::centml"],
+        "memory": ["inline::faiss", "remote::chromadb", "remote::pgvector"],
+        "safety": ["inline::llama-guard"],
+        "agents": ["inline::meta-reference"],
+        "telemetry": ["inline::meta-reference"],
+        "eval": ["inline::meta-reference"],
+        "datasetio": ["remote::huggingface", "inline::localfs"],
+        "scoring": ["inline::basic", "inline::llm-as-judge", "inline::braintrust"],
+    }
+    name = "centml"
+
+    # Primary inference provider: CentML
+    inference_provider = Provider(
+        provider_id="centml",
+        provider_type="remote::centml",
+        config=CentMLImplConfig.sample_run_config(),
+    )
+
+    # Memory provider: Faiss
+    memory_provider = Provider(
+        provider_id="faiss",
+        provider_type="inline::faiss",
+        config=FaissImplConfig.sample_run_config(f"distributions/{name}"),
+    )
+
+    # Embedding provider: SentenceTransformers
+    embedding_provider = Provider(
+        provider_id="sentence-transformers",
+        provider_type="inline::sentence-transformers",
+        config=SentenceTransformersInferenceConfig.sample_run_config(),
+    )
+
+    # Map Llama Models to provider IDs if needed
+    core_model_to_hf_repo = {
+        m.descriptor(): m.huggingface_repo for m in all_registered_models()
+    }
+    default_models = [
+        ModelInput(
+            model_id=core_model_to_hf_repo[m.llama_model],
+            provider_model_id=m.provider_model_id,
+            provider_id="centml",
+        )
+        for m in MODEL_ALIASES
+    ]
+
+    # Example embedding model
+    embedding_model = ModelInput(
+        model_id="all-MiniLM-L6-v2",
+        provider_id="sentence-transformers",
+        model_type=ModelType.embedding,
+        metadata={"embedding_dimension": 384},
+    )
+
+    return DistributionTemplate(
+        name=name,
+        distro_type="self_hosted",
+        description="Use CentML for running LLM inference",
+        docker_image=None,
+        template_path=Path(__file__).parent / "doc_template.md",
+        providers=providers,
+        default_models=default_models,
+        run_configs={
+            "run.yaml": RunConfigSettings(
+                provider_overrides={
+                    "inference": [inference_provider, embedding_provider],
+                    "memory": [memory_provider],
+                },
+                default_models=default_models + [embedding_model],
+                default_shields=[ShieldInput(shield_id="meta-llama/Llama-Guard-3-8B")],
+            ),
+        },
+        run_config_env_vars={
+            "LLAMASTACK_PORT": (
+                "5001",
+                "Port for the Llama Stack distribution server",
+            ),
+            "CENTML_API_KEY": (
+                "",
+                "CentML API Key",
+            ),
+        },
+    )

llama_stack/templates/centml/doc_template.md

@@ -0,0 +1,66 @@
+---
+orphan: true
+---
+# CentML Distribution
+
+```{toctree}
+:maxdepth: 2
+:hidden:
+
+self
+```
+
+The `llamastack/distribution-{{ name }}` distribution consists of the following provider configurations.
+
+{{ providers_table }}
+
+{% if run_config_env_vars %}
+### Environment Variables
+
+The following environment variables can be configured:
+
+{% for var, (default_value, description) in run_config_env_vars.items() %}
+- `{{ var }}`: {{ description }} (default: `{{ default_value }}`)
+{% endfor %}
+{% endif %}
+
+{% if default_models %}
+### Models
+
+The following models are available by default:
+
+{% for model in default_models %}
+- `{{ model.model_id }}`
+{% endfor %}
+{% endif %}
+
+### Prerequisite: API Keys
+
+Make sure you have a valid **CentML API Key**. Sign up or access your credentials at [CentML.com](https://centml.com/).
+
+## Running Llama Stack with CentML
+
+You can do this via Conda (build code) or Docker which has a pre-built image.
+
+### Via Docker
+
+This method allows you to get started quickly without having to build the distribution code.
+
+```bash
+LLAMA_STACK_PORT=5001
+docker run \
+  -it \
+  -p $LLAMA_STACK_PORT:$LLAMA_STACK_PORT \
+  llamastack/distribution-{{ name }} \
+  --port $LLAMA_STACK_PORT \
+  --env CENTML_API_KEY=$CENTML_API_KEY
+```
+
+### Via Conda
+
+```bash
+llama stack build --template {{ name }} --image-type conda
+llama stack run ./run.yaml \
+  --port $LLAMA_STACK_PORT \
+  --env CENTML_API_KEY=$CENTML_API_KEY
+```

llama_stack/templates/centml/run.yaml

@@ -0,0 +1,129 @@
+version: '2'
+image_name: centml
+conda_env: centml
+apis:
+  - agents
+  - datasetio
+  - eval
+  - inference
+  - memory
+  - safety
+  - scoring
+  - telemetry
+  - tool_runtime
+providers:
+  inference:
+  - provider_id: centml
+    provider_type: remote::centml
+    config:
+      url: https://api.centml.com/openai/v1
+      api_key: "${env.CENTML_API_KEY}"
+  - provider_id: sentence-transformers
+    provider_type: inline::sentence-transformers
+    config: {}
+
+  memory:
+  - provider_id: faiss
+    provider_type: inline::faiss
+    config:
+      kvstore:
+        type: sqlite
+        namespace: null
+        db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/centml}/faiss_store.db
+
+  safety:
+  - provider_id: llama-guard
+    provider_type: inline::llama-guard
+    config: {}
+
+  agents:
+  - provider_id: meta-reference
+    provider_type: inline::meta-reference
+    config:
+      persistence_store:
+        type: sqlite
+        namespace: null
+        db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/centml}/agents_store.db
+
+  telemetry:
+  - provider_id: meta-reference
+    provider_type: inline::meta-reference
+    config:
+      service_name: ${env.OTEL_SERVICE_NAME:llama-stack}
+      sinks: ${env.TELEMETRY_SINKS:console,sqlite}
+      sqlite_db_path: ${env.SQLITE_DB_PATH:~/.llama/distributions/centml}/trace_store.db
+
+  eval:
+  - provider_id: meta-reference
+    provider_type: inline::meta-reference
+    config: {}
+
+  datasetio:
+  - provider_id: huggingface
+    provider_type: remote::huggingface
+    config: {}
+  - provider_id: localfs
+    provider_type: inline::localfs
+    config: {}
+
+  scoring:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: llm-as-judge
+    provider_type: inline::llm-as-judge
+    config: {}
+  - provider_id: braintrust
+    provider_type: inline::braintrust
+    config:
+      openai_api_key: ${env.OPENAI_API_KEY:}
+
+  tool_runtime:
+  - provider_id: brave-search
+    provider_type: remote::brave-search
+    config:
+      api_key: ${env.BRAVE_SEARCH_API_KEY:}
+      max_results: 3
+  - provider_id: tavily-search
+    provider_type: remote::tavily-search
+    config:
+      api_key: ${env.TAVILY_SEARCH_API_KEY:}
+      max_results: 3
+  - provider_id: code-interpreter
+    provider_type: inline::code-interpreter
+    config: {}
+  - provider_id: memory-runtime
+    provider_type: inline::memory-runtime
+    config: {}
+
+metadata_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/centml}/registry.db
+
+models:
+- metadata: {}
+  model_id: meta-llama/Llama-3.3-70B-Instruct
+  provider_id: centml
+  provider_model_id: meta-llama/Llama-3.3-70B-Instruct
+  model_type: llm
+
+- metadata: {}
+  model_id: meta-llama/Llama-3.1-405B-Instruct-FP8
+  provider_id: centml
+  provider_model_id: meta-llama/Llama-3.1-405B-Instruct-FP8
+  model_type: llm
+
+shields:
+  - shield_id: meta-llama/Llama-Guard-3-8B
+
+memory_banks: []
+datasets: []
+scoring_fns: []
+eval_tasks: []
+tool_groups:
+  - toolgroup_id: builtin::websearch
+    provider_id: tavily-search
+  - toolgroup_id: builtin::memory
+    provider_id: memory-runtime
+  - toolgroup_id: builtin::code_interpreter
+    provider_id: code-interpreter