Merge branch 'main' into hide-collection-types

microsoft · Oct 13, 2023 · d4024ab · d4024ab
2 parents d8fee5d + 988f65e
commit d4024ab
Show file tree

Hide file tree

Showing 4 changed files with 20 additions and 13 deletions.
diff --git a/python/semantic_kernel/connectors/ai/hugging_face/services/hf_text_completion.py b/python/semantic_kernel/connectors/ai/hugging_face/services/hf_text_completion.py
@@ -23,7 +23,7 @@ class HuggingFaceTextCompletion(TextCompletionClientBase):
     def __init__(
         self,
         model_id: str,
-        device: Optional[int] = -1,
+        device: Optional[int] = None,
         task: Optional[str] = None,
         log: Optional[Logger] = None,
         model_kwargs: Dict[str, Any] = None,
@@ -35,7 +35,10 @@ def __init__(
         Arguments:
             model_id {str} -- Hugging Face model card string, see
                 https://huggingface.co/models
-            device {Optional[int]} -- Device to run the model on, -1 for CPU, 0+ for GPU.
+            device {Optional[int]} -- Device to run the model on, defaults to CPU, 0+ for GPU,
+                                   -- None if using device_map instead. (If both device and device_map
+                                      are specified, device overrides device_map. If unintended,
+                                      it can lead to unexpected behavior.)
             task {Optional[str]} -- Model completion task type, options are:
                 - summarization: takes a long text and returns a shorter summary.
                 - text-generation: takes incomplete text and returns a set of completion candidates.
@@ -64,11 +67,15 @@ def __init__(
                 "Please ensure that torch and transformers are installed to use HuggingFaceTextCompletion"
             )
 
-        self.device = (
-            "cuda:" + str(device)
-            if device >= 0 and torch.cuda.is_available()
-            else "cpu"
-        )
+        device_map = self._pipeline_kwargs.get("device_map", None)
+        if device is None:
+            self.device = "cpu" if device_map is None else None
+        else:
+            self.device = (
+                "cuda:" + str(device)
+                if device >= 0 and torch.cuda.is_available()
+                else "cpu"
+            )
 
         self.generator = transformers.pipeline(
             task=self._task,

diff --git a/python/semantic_kernel/connectors/ai/open_ai/services/open_ai_chat_completion.py b/python/semantic_kernel/connectors/ai/open_ai/services/open_ai_chat_completion.py
@@ -63,7 +63,7 @@ def __init__(
         self._org_id = org_id
         self._api_type = api_type
         self._api_version = api_version
-        self._endpoint = endpoint
+        self._endpoint = endpoint.rstrip("/") if endpoint is not None else None
         self._log = log if log is not None else NullLogger()
         self._messages = []
 
@@ -271,7 +271,7 @@ async def _send_chat_request(
         except Exception as ex:
             raise AIException(
                 AIException.ErrorCodes.ServiceError,
-                "OpenAI service failed to complete the chat",
+                f"{self.__class__.__name__} failed to complete the chat",
                 ex,
             ) from ex
 

diff --git a/python/semantic_kernel/connectors/ai/open_ai/services/open_ai_text_completion.py b/python/semantic_kernel/connectors/ai/open_ai/services/open_ai_text_completion.py
@@ -53,7 +53,7 @@ def __init__(
         self._api_key = api_key
         self._api_type = api_type
         self._api_version = api_version
-        self._endpoint = endpoint
+        self._endpoint = endpoint.rstrip("/") if endpoint is not None else None
         self._org_id = org_id
         self._log = log if log is not None else NullLogger()
 
@@ -161,7 +161,7 @@ async def _send_completion_request(
         except Exception as ex:
             raise AIException(
                 AIException.ErrorCodes.ServiceError,
-                "OpenAI service failed to complete the prompt",
+                f"{self.__class__.__name__} failed to complete the prompt",
                 ex,
             )
 

diff --git a/python/semantic_kernel/connectors/ai/open_ai/services/open_ai_text_embedding.py b/python/semantic_kernel/connectors/ai/open_ai/services/open_ai_text_embedding.py
@@ -48,7 +48,7 @@ def __init__(
         self._api_key = api_key
         self._api_type = api_type
         self._api_version = api_version
-        self._endpoint = endpoint
+        self._endpoint = endpoint.rstrip("/") if endpoint is not None else None
         self._org_id = org_id
         self._log = log if log is not None else NullLogger()
 
@@ -81,6 +81,6 @@ async def generate_embeddings_async(
         except Exception as ex:
             raise AIException(
                 AIException.ErrorCodes.ServiceError,
-                "OpenAI service failed to generate embeddings",
+                f"{self.__class__.__name__} failed to generate embeddings",
                 ex,
             )