Added more deployments to the docs

aana/deployments/__init__.py

@@ -6,11 +6,20 @@
     LLMBatchOutput,
     LLMOutput,
 )
+from aana.deployments.hf_text_generation_deployment import (
+    HfTextGenerationConfig,
+    HfTextGenerationDeployment,
+)
+from aana.deployments.vllm_deployment import VLLMConfig, VLLMDeployment
 
 __all__ = [
     "AanaDeploymentHandle",
     "BaseDeployment",
     "BaseTextGenerationDeployment",
+    "HfTextGenerationConfig",
+    "HfTextGenerationDeployment",
+    "VLLMConfig",
+    "VLLMDeployment",
     "ChatOutput",
     "LLMBatchOutput",
     "LLMOutput",

aana/deployments/vllm_deployment.py

@@ -32,17 +32,22 @@ class VLLMConfig(BaseModel):
     """The configuration of the vLLM deployment.
 
     Attributes:
-        model (str): the model name
-        dtype (str): the data type (optional, default: "auto")
-        quantization (str): the quantization method (optional, default: None)
-        gpu_memory_reserved (float): the GPU memory reserved for the model in mb
-        default_sampling_params (SamplingParams): the default sampling parameters.
-        max_model_len (int): the maximum generated text length in tokens (optional, default: None)
-        chat_template (str): the name of the chat template, if not provided, the chat template from the model will be used
-                             but some models may not have a chat template (optional, default: None)
-        enforce_eager: whether to enforce eager execution (optional, default: False)
-        engine_args: extra engine arguments (optional, default: {})
-
+        model: The model name.
+        dtype: The data type.
+            Defaults to "auto".
+        quantization: The quantization method.
+            Defaults to None.
+        gpu_memory_reserved: The GPU memory reserved for the model in MB.
+        default_sampling_params: The default sampling parameters.
+        max_model_len: The maximum generated text length in tokens.
+            Defaults to None.
+        chat_template: The name of the chat template. If not provided, the chat template
+            from the model will be used. Some models may not have a chat template.
+            Defaults to None.
+        enforce_eager: Whether to enforce eager execution.
+            Defaults to False.
+        engine_args: Extra engine arguments.
+            Defaults to {}.
     """
 
     model: str

docs/reference/aana_deployments.md

@@ -1,3 +1,32 @@
 # Aana Deployments
 
-::: aana.deployments
+## AanaDeploymentHandle
+
+AanaDeploymentHandle is a class that allows you to interact with Aana deployments.
+
+::: aana.deployments.AanaDeploymentHandle
+
+## Base classes for deployments
+
+At the momemnt there are two base classes that you can use to create your own deployments:
+- BaseDeployment: This is the base class for all Aana deployments.
+- BaseTextGenerationDeployment: This is the base class for all text generation deployments (LLM deployments).
+
+::: aana.deployments.BaseDeployment
+::: aana.deployments.BaseTextGenerationDeployment
+
+## Text generation deployments
+
+### Hugging Face Text Generation Deployment
+
+Hugging Face Text Generation Deployment allows you to use Hugging Face transformers library to serve LLMs.
+
+::: aana.deployments.HfTextGenerationConfig
+::: aana.deployments.HfTextGenerationDeployment
+
+### VLLM Deployment
+
+VLLM Deployment allows you to use vLLM library to serve LLMs.
+
+::: aana.deployments.VLLMConfig
+::: aana.deployments.VLLMDeployment

mkdocs.yml

@@ -25,13 +25,13 @@ theme:
       toggle:
         icon: material/toggle-switch-off-outline 
         name: Switch to dark mode
-      primary: teal
+      primary: 5c21d1
       accent: purple 
     - scheme: slate 
       toggle:
         icon: material/toggle-switch
         name: Switch to light mode    
-      primary: teal
+      primary: 5c21d1
       accent: lime
 
 # Extensions
@@ -52,18 +52,19 @@ markdown_extensions:
       emoji_index: !!python/name:materialx.emoji.twemoji
       emoji_generator: !!python/name:materialx.emoji.to_svg
   - toc:
-      toc_depth: 3
+      toc_depth: 2
 
 # Plugins
 plugins:
   mkdocstrings:
     handlers:
       python:
         options:
+          show_object_full_path: false
           show_root_heading: true
           show_if_no_docstring: true
           show_root_toc_entry: false
-          inherited_members: true 
+          inherited_members: true
           members_order: source
           separate_signature: true
           unwrap_annotated: true
@@ -73,6 +74,7 @@ plugins:
           signature_crossrefs: true
           show_symbol_type_heading: true
           show_symbol_type_toc: true
+          show_labels: false
 # Customization
 extra:
   social: