How to customise the model path for model_1200000 and model-00001-of-00002?

[sparrowtraveler]

Checks

• This template is only for question, not feature requests or bug reports.
• I have thoroughly reviewed the project documentation and read the related paper(s).
• I have searched for existing issues, including closed ones, no similar questions.
• I confirm that I am using English to submit this report in order to facilitate communication.

Question details

How to customise the model path of model, model_1200000 and model-00001-of-00002 in F5-tts, do I need to modify the infer_gradio.py file? Can you tell me the exact way to modify the path code?

[SWivid]

How to customise the model path of model, model_1200000 and model-00001-of-00002 in F5-tts, do I need to modify the infer_gradio.py file? Can you tell me the exact way to modify the path code?

what do you mean by model-00001-of-00002 here?

[sparrowtraveler]

infer_gradio.py file
This model was downloaded when I first started the voice chat, I mean to transfer the model of F5-TTS to other disk to reduce the memory consumption of the system disk, how can I customize the path of the model? Modify the infer_gradio.py file? Can you give me more details?

[SWivid]

ckpt_path=str(...)

it is a str, you could just replace with the local_path

model-00001-of-00002 is the chat model (check into the .from_pretrained() to see how you could replace with loading from local)

F5-TTS/src/f5_tts/infer/infer_gradio.py

Lines 551 to 571 in f992c4e

	if not USING_SPACES:
	load_chat_model_btn = gr.Button("Load Chat Model", variant="primary")
	chat_interface_container = gr.Column(visible=False)
	@gpu_decorator
	def load_chat_model():
	global chat_model_state, chat_tokenizer_state
	if chat_model_state is None:
	show_info = gr.Info
	show_info("Loading chat model...")
	model_name = "Qwen/Qwen2.5-3B-Instruct"
	chat_model_state = AutoModelForCausalLM.from_pretrained(
	model_name, torch_dtype="auto", device_map="auto"
	)
	chat_tokenizer_state = AutoTokenizer.from_pretrained(model_name)
	show_info("Chat model loaded.")
	return gr.update(visible=False), gr.update(visible=True)
	load_chat_model_btn.click(load_chat_model, outputs=[load_chat_model_btn, chat_interface_container])

[sparrowtraveler]

ckpt_path=str(...)

it is a str, you could just replace with the local_path

model-00001-of-00002 is the chat model (check into the .from_pretrained() to see how you could replace with loading from local)

F5-TTS/src/f5_tts/infer/infer_gradio.py

Lines 551 to 571 in f992c4e

	if not USING_SPACES:
	load_chat_model_btn = gr.Button("Load Chat Model", variant="primary")
	chat_interface_container = gr.Column(visible=False)
	@gpu_decorator
	def load_chat_model():
	global chat_model_state, chat_tokenizer_state
	if chat_model_state is None:
	show_info = gr.Info
	show_info("Loading chat model...")
	model_name = "Qwen/Qwen2.5-3B-Instruct"
	chat_model_state = AutoModelForCausalLM.from_pretrained(
	model_name, torch_dtype="auto", device_map="auto"
	)
	chat_tokenizer_state = AutoTokenizer.from_pretrained(model_name)
	show_info("Chat model loaded.")
	return gr.update(visible=False), gr.update(visible=True)
	load_chat_model_btn.click(load_chat_model, outputs=[load_chat_model_btn, chat_interface_container])

Thanks. I'll try it.