From 15926210d3ac729f089ff57f743cbb2a5024047c Mon Sep 17 00:00:00 2001 From: Mohit Sharma Date: Wed, 23 Apr 2025 08:09:51 +0000 Subject: [PATCH] disable chunking for qwen --- server/text_generation_server/models/__init__.py | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/server/text_generation_server/models/__init__.py b/server/text_generation_server/models/__init__.py index 20d02bbe..0a211ec3 100644 --- a/server/text_generation_server/models/__init__.py +++ b/server/text_generation_server/models/__init__.py @@ -1522,6 +1522,8 @@ def get_model( kv_cache_dtype=kv_cache_dtype, trust_remote_code=trust_remote_code, lora_adapter_ids=lora_adapter_ids, + # TODO: Fix bug in rust image_text_replacement implementation + support_chunking=False, ) # TODO: Uncomment when transformers is refactored # elif FLASH_TRANSFORMERS_BACKEND: @@ -1553,6 +1555,8 @@ def get_model( lora_adapter_ids=lora_adapter_ids, config_class=Qwen2_5_VLConfig, processor_class=Qwen2_5_VLProcessor, + # TODO: Fix bug in rust image_text_replacement implementation + support_chunking=False, ) # TODO: Uncomment when transformers is refactored # elif FLASH_TRANSFORMERS_BACKEND: