Support third-party Inference providers in `HfApiModel` (#422)

* Add `provider` param to `HfApiModel `, update guided_tour.md

---------

Co-authored-by: Aymeric <aymeric.roucher@gmail.com>

docs/source/en/guided_tour.md

@@ -25,18 +25,18 @@ To initialize a minimal agent, you need at least these two arguments:
 
 - `model`, a text-generation model to power your agent - because the agent is different from a simple LLM, it is a system that uses a LLM as its engine. You can use any of these options:
     - [`TransformersModel`] takes a pre-initialized `transformers` pipeline to run inference on your local machine using `transformers`.
-    - [`HfApiModel`] leverages a `huggingface_hub.InferenceClient` under the hood.
-    - [`LiteLLMModel`] lets you call 100+ different models through [LiteLLM](https://docs.litellm.ai/)!
+    - [`HfApiModel`] leverages a `huggingface_hub.InferenceClient` under the hood and supports all Inference Providers on the Hub.
+    - [`LiteLLMModel`] similarly lets you call 100+ different models and providers through [LiteLLM](https://docs.litellm.ai/)!
     - [`AzureOpenAIServerModel`] allows you to use OpenAI models deployed in [Azure](https://azure.microsoft.com/en-us/products/ai-services/openai-service).
 
 - `tools`, a list of `Tools` that the agent can use to solve the task. It can be an empty list. You can also add the default toolbox on top of your `tools` list by defining the optional argument `add_base_tools=True`.
 
-Once you have these two arguments, `tools` and `model`,  you can create an agent and run it. You can use any LLM you'd like, either through [Hugging Face API](https://huggingface.co/docs/api-inference/en/index), [transformers](https://github.com/huggingface/transformers/), [ollama](https://ollama.com/), [LiteLLM](https://www.litellm.ai/), or [Azure OpenAI](https://azure.microsoft.com/en-us/products/ai-services/openai-service).
+Once you have these two arguments, `tools` and `model`,  you can create an agent and run it. You can use any LLM you'd like, either through [Inference Providers](https://huggingface.co/blog/inference-providers), [transformers](https://github.com/huggingface/transformers/), [ollama](https://ollama.com/), [LiteLLM](https://www.litellm.ai/), or [Azure OpenAI](https://azure.microsoft.com/en-us/products/ai-services/openai-service).
 
 <hfoptions id="Pick a LLM">
-<hfoption id="Hugging Face API">
+<hfoption id="HF Inference API">
 
-Hugging Face API is free to use without a token, but then it will have a rate limitation.
+HF Inference API is free to use without a token, but then it will have a rate limit.
 
 To access gated models or rise your rate limits with a PRO account, you need to set the environment variable `HF_TOKEN` or pass `token` variable upon initialization of `HfApiModel`. You can get your token from your [settings page](https://huggingface.co/settings/tokens)
 
@@ -46,6 +46,7 @@ from smolagents import CodeAgent, HfApiModel
 model_id = "meta-llama/Llama-3.3-70B-Instruct" 
 
 model = HfApiModel(model_id=model_id, token="<YOUR_HUGGINGFACEHUB_API_TOKEN>") # You can choose to not pass any model_id to HfApiModel to use a default free model
+# you can also specify a particular provider e.g. provider="together" or provider="sambanova"
 agent = CodeAgent(tools=[], model=model, add_base_tools=True)
 
 agent.run(

docs/source/en/index.md

@@ -25,7 +25,7 @@ This library offers:
 
 ✨ **Simplicity**: the logic for agents fits in ~thousand lines of code. We kept abstractions to their minimal shape above raw code!
 
-🌐 **Support for any LLM**: it supports models hosted on the Hub loaded in their `transformers` version or through our inference API, but also models from OpenAI, Anthropic... it's really easy to power an agent with any LLM.
+🌐 **Support for any LLM**: it supports models hosted on the Hub loaded in their `transformers` version or through our inference API and Inference providers, but also models from OpenAI, Anthropic... it's really easy to power an agent with any LLM.
 
 🧑‍💻 **First-class support for Code Agents**, i.e. agents that write their actions in code (as opposed to "agents being used to write code"), [read more here](tutorials/secure_code_execution).
 

docs/source/en/reference/models.md

@@ -74,7 +74,7 @@ print(model([{"role": "user", "content": "Ok!"}], stop_sequences=["great"]))
 
 ### HfApiModel
 
-The `HfApiModel` wraps an [HF Inference API](https://huggingface.co/docs/api-inference/index) client for the execution of the LLM.
+The `HfApiModel` wraps huggingface_hub's [InferenceClient](https://huggingface.co/docs/huggingface_hub/main/en/guides/inference) for the execution of the LLM. It supports both HF's own [Inference API](https://huggingface.co/docs/api-inference/index) as well as all [Inference Providers](https://huggingface.co/blog/inference-providers) available on the Hub.
 
 ```python
 from smolagents import HfApiModel

pyproject.toml

@@ -12,7 +12,7 @@ authors = [
 readme = "README.md"
 requires-python = ">=3.10"
 dependencies = [
-  "huggingface-hub>=0.24.0",
+  "huggingface-hub>=0.28.0",
   "requests>=2.32.3",
   "rich>=13.9.4",
   "pandas>=2.2.3",

src/smolagents/__init__.py

@@ -21,6 +21,8 @@ from .default_tools import *
 from .e2b_executor import *
 from .gradio_ui import *
 from .local_python_executor import *
+from .logger import *
+from .memory import *
 from .models import *
 from .monitoring import *
 from .prompts import *

src/smolagents/models.py

@@ -338,6 +338,9 @@ class HfApiModel(Model):
     Parameters:
         model_id (`str`, *optional*, defaults to `"Qwen/Qwen2.5-Coder-32B-Instruct"`):
             The Hugging Face model ID to be used for inference. This can be a path or model identifier from the Hugging Face model hub.
+        provider (`str`, *optional*):
+            Name of the provider to use for inference. Can be `"replicate"`, `"together"`, `"fal-ai"`, `"sambanova"` or `"hf-inference"`.
+            defaults to hf-inference (HF Inference API).
         token (`str`, *optional*):
             Token used by the Hugging Face API for authentication. This token need to be authorized 'Make calls to the serverless Inference API'.
             If the model is gated (like Llama-3 models), the token also needs 'Read access to contents of all public gated repos you can access'.
@@ -368,15 +371,17 @@ class HfApiModel(Model):
     def __init__(
         self,
         model_id: str = "Qwen/Qwen2.5-Coder-32B-Instruct",
+        provider: Optional[str] = None,
         token: Optional[str] = None,
         timeout: Optional[int] = 120,
         **kwargs,
     ):
         super().__init__(**kwargs)
         self.model_id = model_id
+        self.provider = provider
         if token is None:
             token = os.getenv("HF_TOKEN")
-        self.client = InferenceClient(self.model_id, token=token, timeout=timeout)
+        self.client = InferenceClient(self.model_id, provider=provider, token=token, timeout=timeout)
 
     def __call__(
         self,

tests/test_agents.py

@@ -474,7 +474,7 @@ class AgentTests(unittest.TestCase):
         with agent.logger.console.capture() as capture:
             agent.run("Count to 3")
         str_output = capture.get()
-        assert "Consider passing said import under" in str_output.replace("\n", "")
+        assert "`additional_authorized_imports`" in str_output.replace("\n", "")
 
     def test_multiagents(self):
         class FakeModelMultiagentsManagerAgent:

tests/test_all_docs.py

@@ -78,6 +78,7 @@ class DocCodeExtractor:
         return tmp_file
 
 
+@pytest.mark.skipif(not os.getenv("RUN_ALL"), reason="RUN_ALL environment variable not set")
 class TestDocs:
     """Test case for documentation code testing."""
 

tests/test_models.py

@@ -13,10 +13,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 import json
+import os
 import unittest
 from pathlib import Path
 from typing import Optional
 
+import pytest
 from transformers.testing_utils import get_tests_dir
 
 from smolagents import ChatMessage, HfApiModel, TransformersModel, models, tool
@@ -51,6 +53,12 @@ class ModelTests(unittest.TestCase):
         messages = [{"role": "user", "content": [{"type": "text", "text": "Hello!"}]}]
         model(messages, stop_sequences=["great"])
 
+    @pytest.mark.skipif(not os.getenv("RUN_ALL"), reason="RUN_ALL environment variable not set")
+    def test_get_hfapi_message_no_tool_external_provider(self):
+        model = HfApiModel(model="Qwen/Qwen2.5-Coder-32B-Instruct", provider="together", max_tokens=10)
+        messages = [{"role": "user", "content": [{"type": "text", "text": "Hello!"}]}]
+        model(messages, stop_sequences=["great"])
+
     def test_transformers_message_no_tool(self):
         model = TransformersModel(
             model_id="HuggingFaceTB/SmolLM2-135M-Instruct",