From 45b966363ec78d5a667a9b38cb571e4c2da6da37 Mon Sep 17 00:00:00 2001
From: DocShotgun <126566557+DocShotgun@users.noreply.github.com>
Date: Sat, 3 May 2025 21:01:03 -0700
Subject: [PATCH] Tree: Format

---
 backends/exllamav3/model.py            | 12 +++++++-----
 endpoints/OAI/types/chat_completion.py |  2 +-
 2 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/backends/exllamav3/model.py b/backends/exllamav3/model.py
index 1026faf..61986c1 100644
--- a/backends/exllamav3/model.py
+++ b/backends/exllamav3/model.py
@@ -165,9 +165,7 @@ class ExllamaV3Container(BaseModelContainer):
             self.draft_model_dir = draft_model_path
             self.draft_config = Config.from_directory(str(draft_model_path.resolve()))
             self.draft_model = Model.from_config(self.draft_config)
-            logger.info(
-                f'Using draft model: {str(draft_model_path.resolve())}'
-            )
+            logger.info(f"Using draft model: {str(draft_model_path.resolve())}")
         else:
             self.draft_model = None
             self.craft_cache = None
@@ -262,7 +260,9 @@ class ExllamaV3Container(BaseModelContainer):
                 case "Q8":
                     self.draft_cache_mode = "8,8"
 
-            split_draft_cache_mode = re.search(r"^([2-8])\s*,\s*([2-8])$", self.draft_cache_mode)
+            split_draft_cache_mode = re.search(
+                r"^([2-8])\s*,\s*([2-8])$", self.draft_cache_mode
+            )
             if split_draft_cache_mode:
                 draft_k_bits = int(split_draft_cache_mode.group(1))
                 draft_v_bits = int(split_draft_cache_mode.group(2))
@@ -274,7 +274,9 @@ class ExllamaV3Container(BaseModelContainer):
                     v_bits=draft_v_bits,
                 )
             else:
-                self.draft_cache = Cache(self.draft_model, max_num_tokens = self.cache_size)
+                self.draft_cache = Cache(
+                    self.draft_model, max_num_tokens=self.cache_size
+                )
 
         # Max batch size
         self.max_batch_size = unwrap(kwargs.get("max_batch_size"), 256)
diff --git a/endpoints/OAI/types/chat_completion.py b/endpoints/OAI/types/chat_completion.py
index 51695c2..fb73eb9 100644
--- a/endpoints/OAI/types/chat_completion.py
+++ b/endpoints/OAI/types/chat_completion.py
@@ -84,7 +84,7 @@ class ChatCompletionRequest(CommonCompletionRequest):
 
     # Chat completions requests do not have a BOS token preference. Backend
     # respects the tokenization config for the individual model.
-    add_bos_token: Optional[bool] = Field(default = None)
+    add_bos_token: Optional[bool] = Field(default=None)
 
     @field_validator("add_bos_token", mode="after")
     def force_bos_token(cls, v):