From 51f3af11b57995a98d4f7c1fb06ddd5d240e0e22 Mon Sep 17 00:00:00 2001 From: Debanjum Date: Sun, 19 Jan 2025 12:26:00 +0700 Subject: [PATCH] Fix Qwen 2.5 14B model source to use Q4_K_M quantized model The official Qwen2.5 14B model doesn't mention standard quantization suffixes like Q4_K_M, so doesn't work with Khoj --- src/khoj/processor/conversation/utils.py | 2 +- src/khoj/utils/constants.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/khoj/processor/conversation/utils.py b/src/khoj/processor/conversation/utils.py index e8c78c75..75a76918 100644 --- a/src/khoj/processor/conversation/utils.py +++ b/src/khoj/processor/conversation/utils.py @@ -62,7 +62,7 @@ model_to_prompt_size = { "claude-3-5-sonnet-20241022": 60000, "claude-3-5-haiku-20241022": 60000, # Offline Models - "Qwen/Qwen2.5-14B-Instruct-GGUF": 20000, + "bartowski/Qwen2.5-14B-Instruct-GGUF": 20000, "bartowski/Meta-Llama-3.1-8B-Instruct-GGUF": 20000, "bartowski/Llama-3.2-3B-Instruct-GGUF": 20000, "bartowski/gemma-2-9b-it-GGUF": 6000, diff --git a/src/khoj/utils/constants.py b/src/khoj/utils/constants.py index 59534895..6e89594d 100644 --- a/src/khoj/utils/constants.py +++ b/src/khoj/utils/constants.py @@ -15,7 +15,7 @@ default_offline_chat_models = [ "bartowski/Llama-3.2-3B-Instruct-GGUF", "bartowski/gemma-2-9b-it-GGUF", "bartowski/gemma-2-2b-it-GGUF", - "Qwen/Qwen2.5-14B-Instruct-GGUF", + "bartowski/Qwen2.5-14B-Instruct-GGUF", ] default_openai_chat_models = ["gpt-4o-mini", "gpt-4o"] default_gemini_chat_models = ["gemini-1.5-flash", "gemini-1.5-pro"]