|
import spaces
|
|
import gradio as gr
|
|
from joycaption import stream_chat_mod, get_text_model, change_text_model, get_repo_gguf
|
|
|
|
JC_TITLE_MD = "<h1><center>JoyCaption Alpha One Mod</center></h1>"
|
|
JC_DESC_MD = """This space is mod of [fancyfeast/joy-caption-alpha-one](https://huggingface.co/spaces/fancyfeast/joy-caption-alpha-one),
|
|
[Wi-zz/joy-caption-pre-alpha](https://huggingface.co/Wi-zz/joy-caption-pre-alpha). Thanks to [dominic1021](https://huggingface.co/dominic1021)"""
|
|
|
|
css = """
|
|
.info {text-align:center; !important}
|
|
"""
|
|
|
|
with gr.Blocks(fill_width=True, css=css, delete_cache=(60, 3600)) as demo:
|
|
gr.HTML(JC_TITLE_MD)
|
|
with gr.Row():
|
|
with gr.Column():
|
|
with gr.Group():
|
|
jc_input_image = gr.Image(type="pil", label="Input Image", sources=["upload", "clipboard"], height=384)
|
|
with gr.Row():
|
|
jc_caption_type = gr.Dropdown(
|
|
choices=["descriptive", "training_prompt", "rng-tags"],
|
|
label="Caption Type",
|
|
value="descriptive",
|
|
)
|
|
jc_caption_tone = gr.Dropdown(
|
|
choices=["formal", "informal"],
|
|
label="Caption Tone",
|
|
value="formal",
|
|
)
|
|
jc_caption_length = gr.Dropdown(
|
|
choices=["any", "very short", "short", "medium-length", "long", "very long"] +
|
|
[str(i) for i in range(20, 261, 10)],
|
|
label="Caption Length",
|
|
value="any",
|
|
)
|
|
gr.Markdown("**Note:** Caption tone doesn't affect `rng-tags` and `training_prompt`.", elem_classes="info")
|
|
with gr.Accordion("Advanced", open=False):
|
|
with gr.Row():
|
|
jc_text_model = gr.Dropdown(label="LLM Model", info="You can enter a huggingface model repo_id to want to use.",
|
|
choices=get_text_model(), value=get_text_model()[0],
|
|
allow_custom_value=True, interactive=True, min_width=320)
|
|
jc_gguf = gr.Dropdown(label=f"GGUF Filename", choices=[], value="",
|
|
allow_custom_value=True, min_width=320, visible=False)
|
|
jc_nf4 = gr.Checkbox(label="Use NF4 quantization", value=True)
|
|
jc_text_model_button = gr.Button("Load Model", variant="secondary")
|
|
jc_use_inference_client = gr.Checkbox(label="Use Inference Client", value=False, visible=False)
|
|
with gr.Row():
|
|
jc_tokens = gr.Slider(minimum=1, maximum=4096, value=300, step=1, label="Max tokens")
|
|
jc_temperature = gr.Slider(minimum=0.1, maximum=4.0, value=0.6, step=0.1, label="Temperature")
|
|
jc_topp = gr.Slider(minimum=0, maximum=2.0, value=0.9, step=0.01, label="Top-P")
|
|
jc_run_button = gr.Button("Caption", variant="primary")
|
|
with gr.Column():
|
|
jc_output_caption = gr.Textbox(label="Caption", show_copy_button=True)
|
|
gr.Markdown(JC_DESC_MD, elem_classes="info")
|
|
gr.LoginButton()
|
|
gr.DuplicateButton(value="Duplicate Space for private use (This demo does not work on CPU. Requires GPU Space)")
|
|
|
|
jc_run_button.click(fn=stream_chat_mod, inputs=[jc_input_image, jc_caption_type, jc_caption_tone, jc_caption_length,
|
|
jc_tokens, jc_topp, jc_temperature, jc_text_model], outputs=[jc_output_caption])
|
|
jc_text_model_button.click(change_text_model, [jc_text_model, jc_use_inference_client, jc_gguf, jc_nf4], [jc_text_model], show_api=False)
|
|
|
|
jc_use_inference_client.change(change_text_model, [jc_text_model, jc_use_inference_client], [jc_text_model], show_api=False)
|
|
|
|
if __name__ == "__main__":
|
|
|
|
demo.launch()
|
|
|
