Merge branch 'main' of https://github.com/abetlen/llama-cpp-python into main

llama_cpp/_internals.py

@@ -51,6 +51,9 @@ class _LlamaModel:
             self.path_model.encode("utf-8"), self.params
         )
 
+        if self.model is None:
+            raise ValueError(f"Failed to load model from file: {path_model}")
+
     def __del__(self):
         if self.model is not None and self._llama_free_model is not None:
             self._llama_free_model(self.model)
@@ -258,6 +261,9 @@ class _LlamaContext:
             self.model.model, self.params
         )
 
+        if self.ctx is None:
+            raise ValueError("Failed to create llama_context")
+
     def __del__(self):
         if self.ctx is not None and self._llama_free is not None:
             self._llama_free(self.ctx)

llama_cpp/llama.py

@@ -1885,8 +1885,9 @@ class Llama:
         cls,
         repo_id: str,
         filename: Optional[str],
-        local_dir: Optional[Union[str, os.PathLike[str]]] = ".",
+        local_dir: Optional[Union[str, os.PathLike[str]]] = None,
         local_dir_use_symlinks: Union[bool, Literal["auto"]] = "auto",
+        cache_dir: Optional[Union[str, os.PathLike[str]]] = None,
         **kwargs: Any,
     ) -> "Llama":
         """Create a Llama model from a pretrained model name or path.
@@ -1945,17 +1946,28 @@ class Llama:
         subfolder = str(Path(matching_file).parent)
         filename = Path(matching_file).name
 
-        local_dir = "."
-
         # download the file
         hf_hub_download(
             repo_id=repo_id,
-            local_dir=local_dir,
             filename=filename,
             subfolder=subfolder,
+            local_dir=local_dir,
             local_dir_use_symlinks=local_dir_use_symlinks,
+            cache_dir=cache_dir,
         )
 
+        if local_dir is None:
+            model_path = hf_hub_download(
+                repo_id=repo_id,
+                filename=filename,
+                subfolder=subfolder,
+                local_dir=local_dir,
+                local_dir_use_symlinks=local_dir_use_symlinks,
+                cache_dir=cache_dir,
+                local_files_only=True,
+
+            )
+        else:
             model_path = os.path.join(local_dir, filename)
 
         return cls(

llama_cpp/llama_grammar.py

@@ -1498,9 +1498,21 @@ class SchemaConverter:
             item_rule_name = self.visit(
                 schema["items"], f'{name}{"-" if name else ""}item'
             )
-            rule = (
-                f'"[" space ({item_rule_name} ("," space {item_rule_name})*)? "]" space'
-            )
+            list_item_operator = f'("," space {item_rule_name})'
+            successive_items = ""
+            min_items = schema.get("minItems", 0)
+            if min_items > 0:
+               first_item = f"({item_rule_name})"
+               successive_items = list_item_operator * (min_items - 1)
+               min_items -= 1
+            else:
+               first_item = f"({item_rule_name})?"
+            max_items = schema.get("maxItems")
+            if max_items is not None and max_items > min_items:
+                successive_items += (list_item_operator + "?") * (max_items - min_items - 1)
+            else:
+                successive_items += list_item_operator + "*"
+            rule = f'"[" space {first_item} {successive_items} "]" space'
             return self._add_rule(rule_name, rule)
 
         else: