neuralnode 2.0.6__tar.gz → 2.0.8__tar.gz

{neuralnode-2.0.6 → neuralnode-2.0.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: neuralnode
-Version: 2.0.6
+Version: 2.0.8
 Summary: Comprehensive AI Framework with 50+ LLM Providers, Advanced Agents, Chains, Memory, RAG, and 100+ Tools
 Project-URL: Homepage, https://assem.cloud/
 Project-URL: Documentation, https://neuralnode.readthedocs.io

{neuralnode-2.0.6 → neuralnode-2.0.8}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "neuralnode"
-version = "2.0.6"
+version = "2.0.8"
 description = "Comprehensive AI Framework with 50+ LLM Providers, Advanced Agents, Chains, Memory, RAG, and 100+ Tools"
 readme = "README.md"
 requires-python = ">=3.9"

{neuralnode-2.0.6 → neuralnode-2.0.8}/src/neuralnode/__init__.py

@@ -42,7 +42,7 @@ Quick Start::
     text = sr.listen()
 """
 
-__version__ = "2.0.6"
+__version__ = "2.0.8"
 __author__ = "NeuralNode Contributors"
 
 # ── Core types ────────────────────────────────────────────────────────────────
@@ -82,6 +82,7 @@ from .providers import (
     TransformersProvider,
     UniversalLocalProvider,
     # Custom models
+    Horus,
     HorusModel,
     HorusProvider,
 )
@@ -647,7 +648,7 @@ __all__ = [
     "OllamaProvider", "AnthropicProvider", "GoogleProvider",
     "GroqProvider", "MistralProvider", "CohereProvider", "DeepSeekProvider",
     "LlamaCppProvider", "TransformersProvider", "UniversalLocalProvider",
-    "HorusModel", "HorusProvider",
+    "Horus", "HorusModel", "HorusProvider",
 
     # Memory
     "ConversationMemory", "FileMemory", "SlidingWindowMemory", "SummaryMemory",

{neuralnode-2.0.6 → neuralnode-2.0.8}/src/neuralnode/providers/__init__.py

@@ -108,10 +108,11 @@ from .universal_local import UniversalLocalProvider, load_local_model
 
 # ── Custom Model Providers ─────────────────────────────────────────────────────
 try:
-    from .horus import HorusProvider, HorusModel, load_horus
+    from .horus import HorusProvider, HorusModel, Horus, load_horus
 except ImportError:
     HorusProvider = None  # type: ignore
     HorusModel = None  # type: ignore
+    Horus = None  # type: ignore
     load_horus = None  # type: ignore
 
 # MLX Provider (Apple Silicon — optional)
@@ -234,5 +235,5 @@ __all__ = [
     "UniversalLocalProvider", "load_local_model",
     
     # Custom Models
-    "HorusProvider", "HorusModel", "load_horus",
+    "HorusProvider", "HorusModel", "Horus", "load_horus",
 ]

{neuralnode-2.0.6 → neuralnode-2.0.8}/src/neuralnode/providers/horus.py

@@ -141,49 +141,56 @@ class HorusProvider(BaseLLMProvider):
 
     HORUS_MODELS: Dict[str, Dict[str, Any]] = {
         "tokenaii/horus": {
-            "name": "Horus 1.0 4B (Auto)",
+            "name": "Horus-1.0-4B",
+            "official_name": "Horus-1.0-4B",
             "size": "4B",
             "type": "safetensors",
             "repo_id": "tokenaii/horus",
             "subfolder": "Horus-1.0-4B",
         },
         "tokenaii/horus/Horus-1.0-4B": {
-            "name": "Horus 1.0 4B (Full)",
+            "name": "Horus-1.0-4B",
+            "official_name": "Horus-1.0-4B",
             "size": "4B",
             "type": "safetensors",
             "repo_id": "tokenaii/horus",
             "subfolder": "Horus-1.0-4B",
         },
         "tokenaii/Hours-1.0-4B-GGUF/Horus-1.0-4B-Q4_K_M.gguf": {
-            "name": "Horus 1.0 4B (GGUF 4-bit)",
+            "name": "Horus-1.0-4B-Q4_K_M.gguf",
+            "official_name": "Horus-1.0-4B-Q4_K_M.gguf",
             "size": "4B",
             "type": "gguf",
             "quantization": "Q4_K_M",
             "file_size": "2.78 GB",
         },
         "tokenaii/Hours-1.0-4B-GGUF/Horus-1.0-4B-Q5_K_M.gguf": {
-            "name": "Horus 1.0 4B (GGUF 5-bit)",
+            "name": "Horus-1.0-4B-Q5_K_M.gguf",
+            "official_name": "Horus-1.0-4B-Q5_K_M.gguf",
             "size": "4B",
             "type": "gguf",
             "quantization": "Q5_K_M",
             "file_size": "3.23 GB",
         },
         "tokenaii/Hours-1.0-4B-GGUF/Horus-1.0-4B-Q6_K.gguf": {
-            "name": "Horus 1.0 4B (GGUF 6-bit)",
+            "name": "Horus-1.0-4B-Q6_K.gguf",
+            "official_name": "Horus-1.0-4B-Q6_K.gguf",
             "size": "4B",
             "type": "gguf",
             "quantization": "Q6_K",
             "file_size": "3.71 GB",
         },
         "tokenaii/Hours-1.0-4B-GGUF/Horus-1.0-4B-Q8_0.gguf": {
-            "name": "Horus 1.0 4B (GGUF 8-bit)",
+            "name": "Horus-1.0-4B-Q8_0.gguf",
+            "official_name": "Horus-1.0-4B-Q8_0.gguf",
             "size": "4B",
             "type": "gguf",
             "quantization": "Q8_0",
             "file_size": "4.8 GB",
         },
         "tokenaii/Hours-1.0-4B-GGUF/Horus-1.0-4B-F16.gguf": {
-            "name": "Horus 1.0 4B (GGUF F16)",
+            "name": "Horus-1.0-4B-F16.gguf",
+            "official_name": "Horus-1.0-4B-F16.gguf",
             "size": "4B",
             "type": "gguf",
             "quantization": "F16",
@@ -213,6 +220,7 @@ class HorusProvider(BaseLLMProvider):
         self,
         model_id: str = "tokenaii/horus",
         device: Optional[str] = None,
+        n_gpu_layers: Optional[int] = None,
         torch_dtype: Optional[Any] = None,
         load_in_8bit: bool = False,
         load_in_4bit: bool = False,
@@ -276,6 +284,7 @@ class HorusProvider(BaseLLMProvider):
 
         self.model_id = model_id
         self.device = device or self._resolve_device()
+        self.n_gpu_layers = n_gpu_layers
         self.torch_dtype = self._resolve_torch_dtype(torch_dtype)
         self.load_in_8bit = load_in_8bit
         self.load_in_4bit = load_in_4bit
@@ -395,6 +404,7 @@ class HorusProvider(BaseLLMProvider):
     def _build_model_info(self) -> Dict[str, Any]:
         base = dict(self.HORUS_MODELS.get(self.model_id, {}))
         base.setdefault("name", self.model_id.split("/")[-1])
+        base.setdefault("official_name", self._infer_official_name(self.model_id))
         base.setdefault("size", "Unknown")
         base.setdefault("type", "gguf" if self._is_gguf_model_id(self.model_id) else "safetensors")
         if base.get("type") == "gguf" and "quantization" not in base:
@@ -456,6 +466,7 @@ class HorusProvider(BaseLLMProvider):
             "model_path": model_path,
             "n_ctx": HORUS_CONTEXT_WINDOW,
             "verbose": False,
+            "n_gpu_layers": self._resolve_n_gpu_layers(),
         }
         turbo_kwargs = self._resolve_gguf_turboquant_kwargs()
         llama_kwargs.update(turbo_kwargs)
@@ -466,16 +477,19 @@ class HorusProvider(BaseLLMProvider):
                 if turbo_kwargs:
                     self._gguf_kv_quant_applied = True
             except Exception:
-                # Some llama.cpp builds/models reject KV cache dtype overrides.
-                # Fall back to standard GGUF loading without exposing warnings.
-                if turbo_kwargs:
-                    base_kwargs = dict(llama_kwargs)
-                    base_kwargs.pop("type_k", None)
-                    base_kwargs.pop("type_v", None)
+                # Some llama.cpp builds/models reject KV cache dtype overrides
+                # or GPU layer allocation. Fall back progressively without
+                # exposing warnings to end users.
+                base_kwargs = dict(llama_kwargs)
+                base_kwargs.pop("type_k", None)
+                base_kwargs.pop("type_v", None)
+                self._gguf_kv_quant_applied = False
+                try:
                     self.model = Llama(**base_kwargs)
-                    self._gguf_kv_quant_applied = False
-                else:
-                    raise
+                except Exception:
+                    cpu_kwargs = dict(base_kwargs)
+                    cpu_kwargs["n_gpu_layers"] = 0
+                    self.model = Llama(**cpu_kwargs)
         return self
 
     def _resolve_gguf_turboquant_kwargs(self) -> Dict[str, Any]:
@@ -516,6 +530,22 @@ class HorusProvider(BaseLLMProvider):
             result["type_v"] = type_v
         return result
 
+    def _resolve_n_gpu_layers(self) -> int:
+        if self.n_gpu_layers is not None:
+            return int(self.n_gpu_layers)
+        return 0 if self.device.lower() == "cpu" else -1
+
+    @staticmethod
+    def _infer_official_name(model_id: str) -> str:
+        parts = model_id.split("/")
+        if model_id.lower().endswith(".gguf"):
+            return parts[-1]
+        if len(parts) >= 3 and parts[-2] == "horus":
+            return parts[-1]
+        if model_id == "tokenaii/horus":
+            return "Horus-1.0-4B"
+        return parts[-1]
+
     def _load_transformers(self) -> "HorusProvider":
         if not TRANSFORMERS_AVAILABLE and self.auto_install_deps:
             ensure_feature_dependencies("horus_transformers", auto_install=True)
@@ -562,7 +592,6 @@ class HorusProvider(BaseLLMProvider):
             "use_safetensors": self.use_safetensors,
             "proxies": self.proxies,
             "force_download": self.force_download,
-            "resume_download": self.resume_download,
         }
         if self.token:
             model_kwargs["token"] = self.token
@@ -1047,6 +1076,7 @@ class HorusProvider(BaseLLMProvider):
             "id": self.model_id,
             **self._model_info,
             "device": self.device,
+            "n_gpu_layers": self._resolve_n_gpu_layers() if self._is_gguf_model_id(self.model_id) else None,
             "device_map": self.device_map,
             "torch_dtype": str(self.torch_dtype),
             "loaded": self.model is not None,
@@ -1215,6 +1245,7 @@ class HorusProvider(BaseLLMProvider):
         quant = quant_match.group(1).upper() if quant_match else "GGUF"
         cls.HORUS_MODELS[model_id] = {
             "name": name or f"Horus {size} ({quant})",
+            "official_name": filename,
             "size": size,
             "type": "gguf",
             "quantization": quant,
@@ -1295,11 +1326,27 @@ def print_model_list():
 class HorusModel(HorusProvider):
     """Backward-compatible Horus model alias."""
 
-    pass
+    class model:
+        @staticmethod
+        def List() -> List[Dict[str, str]]:
+            """Return available Horus models with exact official Hugging Face names."""
+            items: List[Dict[str, str]] = []
+            for model_id, info in HorusProvider.list_available_models().items():
+                items.append(
+                    {
+                        "model_id": model_id,
+                        "official_name": str(info.get("official_name", HorusProvider._infer_official_name(model_id))),
+                        "type": str(info.get("type", "unknown")),
+                    }
+                )
+            return items
+
+
+Horus = HorusModel
 
 
 def load_horus(model_id: str = "tokenaii/horus", **kwargs) -> HorusProvider:
     return HorusProvider(model_id=model_id, **kwargs).load()
 
 
-__all__ = ["HorusProvider", "HorusModel", "load_horus", "print_model_list"]
+__all__ = ["HorusProvider", "HorusModel", "Horus", "load_horus", "print_model_list"]