Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-19 01:04:32 +02:00
Code Issues Packages Projects Releases Wiki Activity

refactor interrogate/analyze/vqa code

Browse Source 
Signed-off-by: Vladimir Mandic <mandic00@live.com>
This commit is contained in:
Vladimir Mandic
2025-02-01 11:47:20 -05:00
parent ceaf0238fa
commit 654f44f66f
14 changed files with 64 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/ui_postprocessing.py
+15 -14
View File
@@ -1,7 +1,8 @@
import json
import gradio as gr
from modules import scripts, shared, ui_common, postprocessing, call_queue, interrogate, generation_parameters_copypaste
from modules import scripts, shared, ui_common, postprocessing, call_queue, generation_parameters_copypaste
from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call # pylint: disable=unused-import
from modules.interrogate import legacy
def submit_info(image):
@@ -46,8 +47,8 @@ def create_ui():
flavor = gr.Label(elem_id="interrogate_label_flavor", label="Flavor", num_top_classes=5)
with gr.Row():
clip_model = gr.Dropdown([], value='ViT-L-14/openai', label='CLiP model')
ui_common.create_refresh_button(clip_model, interrogate.get_clip_models, lambda: {"choices": interrogate.get_clip_models()}, 'refresh_interrogate_models')
blip_model = gr.Dropdown(list(interrogate.caption_models), value='blip-base', label='Caption model')
ui_common.create_refresh_button(clip_model, legacy.get_clip_models, lambda: {"choices": legacy.get_clip_models()}, 'refresh_interrogate_models')
blip_model = gr.Dropdown(list(legacy.caption_models), value='blip-base', label='Caption model')
mode = gr.Dropdown(['best', 'fast', 'classic', 'caption', 'negative'], label='Mode', value='fast')
with gr.Accordion(label='Advanced', open=False, visible=True):
with gr.Row():
@@ -56,20 +57,20 @@ def create_ui():
min_flavors = gr.Number(label='Min flavors', value=2, minimum=1, maximum=16, min_width=300)
max_flavors = gr.Number(label='Max flavors', value=8, minimum=1, maximum=64, min_width=300)
flavor_intermediate_count = gr.Number(label='Intermediates', value=1024, minimum=256, maximum=4096)
caption_max_length.change(fn=interrogate.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
chunk_size.change(fn=interrogate.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
min_flavors.change(fn=interrogate.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
max_flavors.change(fn=interrogate.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
flavor_intermediate_count.change(fn=interrogate.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
caption_max_length.change(fn=legacy.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
chunk_size.change(fn=legacy.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
min_flavors.change(fn=legacy.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
max_flavors.change(fn=legacy.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
flavor_intermediate_count.change(fn=legacy.update_interrogate_params, inputs=[caption_max_length, chunk_size, min_flavors, max_flavors, flavor_intermediate_count], outputs=[])
with gr.Row(elem_id='interrogate_buttons_image'):
btn_interrogate_img = gr.Button("Interrogate", elem_id="interrogate_btn_interrogate", variant='primary')
btn_analyze_img = gr.Button("Analyze", elem_id="interrogate_btn_analyze", variant='primary')
btn_unload = gr.Button("Unload", elem_id="interrogate_btn_unload")
with gr.Row(elem_id='copy_buttons_interrogate'):
copy_interrogate_buttons = generation_parameters_copypaste.create_buttons(["txt2img", "img2img", "extras", "control"])
btn_interrogate_img.click(interrogate.interrogate_image, inputs=[image, clip_model, blip_model, mode], outputs=prompt)
btn_analyze_img.click(interrogate.analyze_image, inputs=[image, clip_model, blip_model], outputs=[medium, artist, movement, trending, flavor])
btn_unload.click(interrogate.unload_clip_model)
btn_interrogate_img.click(legacy.interrogate_image, inputs=[image, clip_model, blip_model, mode], outputs=prompt)
btn_analyze_img.click(legacy.analyze_image, inputs=[image, clip_model, blip_model], outputs=[medium, artist, movement, trending, flavor])
btn_unload.click(legacy.unload_clip_model)
with gr.Tab("Interrogate Batch"):
with gr.Row():
batch_files = gr.File(label="Files", show_label=True, file_count='multiple', file_types=['image'], type='file', interactive=True, height=100)
@@ -81,11 +82,11 @@ def create_ui():
batch = gr.Text(label="Prompts", lines=10)
with gr.Row():
clip_model = gr.Dropdown([], value='ViT-L-14/openai', label='CLiP Batch Model')
ui_common.create_refresh_button(clip_model, interrogate.get_clip_models, lambda: {"choices": interrogate.get_clip_models()}, 'refresh_interrogate_models')
ui_common.create_refresh_button(clip_model, legacy.get_clip_models, lambda: {"choices": legacy.get_clip_models()}, 'refresh_interrogate_models')
with gr.Row(elem_id='interrogate_buttons_batch'):
btn_interrogate_batch = gr.Button("Interrogate", elem_id="interrogate_btn_interrogate", variant='primary')
with gr.Tab("Visual Query"):
from modules import vqa
from modules.interrogate import vqa
with gr.Row():
vqa_image = gr.Image(type='pil', label="Image")
with gr.Row():
@@ -148,7 +149,7 @@ def create_ui():
]
)
btn_interrogate_batch.click(
fn=interrogate.interrogate_batch,
fn=legacy.interrogate_batch,
inputs=[batch_files, batch_folder, batch_str, clip_model, blip_model, mode, save_output],
outputs=[batch],
)