modelparams 0.0.22__tar.gz → 0.0.23__tar.gz

{modelparams-0.0.22 → modelparams-0.0.23}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: modelparams
-Version: 0.0.22
+Version: 0.0.23
 Summary: Typed model parameters for Python, generated from the modelparams.dev catalog.
 Project-URL: Homepage, https://modelparams.dev
 Project-URL: Repository, https://github.com/mnfst/modelparameters.dev

{modelparams-0.0.22 → modelparams-0.0.23}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "modelparams"
-version = "0.0.22"
+version = "0.0.23"
 description = "Typed model parameters for Python, generated from the modelparams.dev catalog."
 readme = "README.md"
 requires-python = ">=3.10"

{modelparams-0.0.22 → modelparams-0.0.23}/src/modelparams/_generated/catalog.json

@@ -1,4 +1,79 @@
 [
+  {
+    "provider": "alibaba",
+    "authType": "api_key",
+    "model": "deepseek-v4-pro-0813",
+    "params": [
+      {
+        "path": "max_completion_tokens",
+        "label": "Max tokens",
+        "description": "Maximum number of tokens to generate, including both reasoning and the final answer.",
+        "group": "generation_length",
+        "type": "integer",
+        "range": {
+          "min": 1
+        }
+      },
+      {
+        "path": "temperature",
+        "label": "Temperature",
+        "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied.",
+        "group": "sampling",
+        "type": "number",
+        "range": {
+          "min": 0,
+          "max": 1.9,
+          "step": 0.1
+        }
+      },
+      {
+        "path": "top_p",
+        "label": "Top P",
+        "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.",
+        "group": "sampling",
+        "type": "number",
+        "range": {
+          "min": 0,
+          "max": 1,
+          "step": 0.01
+        }
+      },
+      {
+        "path": "extra_body.top_k",
+        "label": "Top K",
+        "description": "Limits generation to the selected number of highest-probability tokens. Values above 100 disable top-k sampling.",
+        "group": "sampling",
+        "type": "integer",
+        "default": 20,
+        "range": {
+          "min": 0
+        }
+      },
+      {
+        "path": "extra_body.enable_thinking",
+        "label": "Enable thinking",
+        "description": "Toggles the model's hybrid thinking mode, sent as a provider-specific extra body field on OpenAI-compatible clients.",
+        "group": "reasoning",
+        "type": "boolean",
+        "default": true
+      },
+      {
+        "path": "extra_body.thinking_budget",
+        "label": "Thinking budget",
+        "description": "Maximum number of tokens the model may spend on reasoning before it starts the final answer; defaults to the model's maximum reasoning length.",
+        "group": "reasoning",
+        "applicability": {
+          "only": {
+            "extra_body.enable_thinking": true
+          }
+        },
+        "type": "integer",
+        "range": {
+          "min": 1
+        }
+      }
+    ]
+  },
   {
     "provider": "alibaba",
     "authType": "api_key",

{modelparams-0.0.22 → modelparams-0.0.23}/src/modelparams/_generated/model_ids.py

@@ -4,6 +4,7 @@
 from typing import Literal
 
 ModelId = Literal[
+    "alibaba/deepseek-v4-pro-0813",
     "alibaba/qwen-flash",
     "alibaba/qwen-max",
     "alibaba/qwen-plus",
@@ -285,6 +286,7 @@ ModelId = Literal[
 ]
 
 MODEL_IDS: tuple[ModelId, ...] = (
+    "alibaba/deepseek-v4-pro-0813",
     "alibaba/qwen-flash",
     "alibaba/qwen-max",
     "alibaba/qwen-plus",

{modelparams-0.0.22 → modelparams-0.0.23}/src/modelparams/_generated/registry.py

@@ -25,6 +25,7 @@ from modelparams.types import z_ai
 from .model_ids import ModelId
 
 PARAM_TYPES: dict[ModelId, Any] = {
+    "alibaba/deepseek-v4-pro-0813": alibaba.Deepseek_V4_Pro_0813Params,
     "alibaba/qwen-flash": alibaba.Qwen_FlashParams,
     "alibaba/qwen-max": alibaba.Qwen_MaxParams,
     "alibaba/qwen-plus": alibaba.Qwen_PlusParams,

{modelparams-0.0.22 → modelparams-0.0.23}/src/modelparams/types/alibaba.py

@@ -10,6 +10,20 @@ from typing_extensions import TypedDict
 
 _PARAMS_CONFIG = ConfigDict(strict=True, extra="forbid")
 
+Deepseek_V4_Pro_0813Params = TypedDict(
+    "Deepseek_V4_Pro_0813Params",
+    {
+        "max_completion_tokens": Annotated[int, Field(ge=1)],
+        "temperature": Annotated[float, Field(ge=0, le=1.9)],
+        "top_p": Annotated[float, Field(ge=0, le=1)],
+        "extra_body.top_k": Annotated[int, Field(ge=0)],
+        "extra_body.enable_thinking": bool,
+        "extra_body.thinking_budget": Annotated[int, Field(ge=1)],
+    },
+    total=False,
+)
+setattr(Deepseek_V4_Pro_0813Params, "__pydantic_config__", _PARAMS_CONFIG)
+
 Qwen_FlashParams = TypedDict(
     "Qwen_FlashParams",
     {
@@ -419,6 +433,7 @@ Qwq_PlusParams = TypedDict(
 setattr(Qwq_PlusParams, "__pydantic_config__", _PARAMS_CONFIG)
 
 __all__ = [
+    "Deepseek_V4_Pro_0813Params",
     "Qwen_FlashParams",
     "Qwen_MaxParams",
     "Qwen_PlusParams",