Use bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf as default model

Author: daavoo

demo/app.py

@@ -23,7 +23,7 @@
 @st.cache_resource
 def load_text_to_text_model():
     return load_llama_cpp_model(
-        model_id="bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf"
+        model_id="bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf"
     )
 
 

demo/notebook.ipynb

@@ -205,7 +205,7 @@
     ")\n",
     "\n",
     "text_model = load_llama_cpp_model(\n",
-    "    \"bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf\"\n",
+    "    \"bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf\"\n",
     ")\n",
     "speech_model = load_tts_model(\"hexgrad/kLegacy/v0.19/kokoro-v0_19.pth\")"
    ]

docs/step-by-step-guide.md

@@ -106,7 +106,7 @@ from document_to_podcast.inference.text_to_text import text_to_text, text_to_tex
 
 # Load the model
 model = load_llama_cpp_model(
-    "bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf"
+    "bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf"
 )
 
 # Define your input and system prompt

example_data/config.yaml

@@ -1,6 +1,6 @@
 input_file: "example_data/introducing-mozilla-ai-investing-in-trustworthy-ai.html"
 output_folder: "example_data/"
-text_to_text_model: "bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf"
+text_to_text_model: "bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf"
 text_to_speech_model: "OuteAI/OuteTTS-0.1-350M-GGUF/OuteTTS-0.1-350M-FP16.gguf"
 text_to_text_prompt: |
   You are a podcast scriptwriter generating engaging and natural-sounding conversations in JSON format. 

src/document_to_podcast/cli.py

@@ -26,7 +26,7 @@
 def document_to_podcast(
     input_file: str | None = None,
     output_folder: str | None = None,
-    text_to_text_model: str = "bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf",
+    text_to_text_model: str = "bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf",
     text_to_text_prompt: str = DEFAULT_PROMPT,
     text_to_speech_model: str = "OuteAI/OuteTTS-0.2-500M-GGUF/OuteTTS-0.2-500M-FP16.gguf",
     speakers: list[Speaker] | None = None,
@@ -58,7 +58,7 @@ def document_to_podcast(
 
             Need to be a gguf file.
 
-            Defaults to `bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf`.
+            Defaults to `bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf`.
 
         text_to_text_prompt (str, optional): The prompt for the text-to-text model.
             Defaults to DEFAULT_PROMPT.

src/document_to_podcast/inference/model_loaders.py

@@ -10,7 +10,7 @@ def load_llama_cpp_model(model_id: str) -> Llama:
     Loads the given model_id using Llama.from_pretrained.
 
     Examples:
-        >>> model = load_llama_cpp_model("bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf")
+        >>> model = load_llama_cpp_model("bartowski/Qwen2.5-7B-Instruct-GGUF/Qwen2.5-7B-Instruct-Q8_0.gguf")
 
     Args:
         model_id (str): The model id to load.