From 0c4cc1eba37768d7b7d4229332a071482a95e500 Mon Sep 17 00:00:00 2001
From: kingbri <8082010+kingbri1@users.noreply.github.com>
Date: Sat, 17 May 2025 21:39:41 -0400
Subject: [PATCH] Model: Add prompt logging to ExllamaV3

Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>
---
 backends/exllamav3/model.py | 7 +++++++
 common/gen_logging.py       | 2 +-
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/backends/exllamav3/model.py b/backends/exllamav3/model.py
index 7b39a33..1fcc8ef 100644
--- a/backends/exllamav3/model.py
+++ b/backends/exllamav3/model.py
@@ -28,6 +28,7 @@ from common.concurrency import iterate_in_threadpool
 from common.gen_logging import (
     log_generation_params,
     log_metrics,
+    log_prompt,
 )
 from common.hardware import hardware_supports_flash_attn
 from common.health import HealthManager
@@ -840,6 +841,12 @@ class ExllamaV3Container(BaseModelContainer):
                 f"max_seq_len {self.max_seq_len}"
             )
 
+        # Log prompt to console. Add the BOS token if specified
+        log_prompt(
+            f"{self.tokenizer.bos_token if add_bos_token else ''}{prompt}",
+            request_id,
+        )
+
         generation = {}
         job = AsyncJob(
             self.generator,
diff --git a/common/gen_logging.py b/common/gen_logging.py
index 150d63c..490d257 100644
--- a/common/gen_logging.py
+++ b/common/gen_logging.py
@@ -29,7 +29,7 @@ def log_generation_params(**kwargs):
         logger.info(f"Generation options: {kwargs}\n")
 
 
-def log_prompt(prompt: str, request_id: str, negative_prompt: Optional[str]):
+def log_prompt(prompt: str, request_id: str, negative_prompt: Optional[str] = None):
     """Logs the prompt to console."""
     if config.logging.log_prompt:
         formatted_prompt = "\n" + prompt