mirror of
https://github.com/vladmandic/automatic
synced 2026-09-18 16:54:33 +02:00
Vladimir Mandic
GitHub
@@ -139,8 +139,6 @@ function requestProgress(id_task, progressbarContainer, gallery, atEnd, onProgre
|
||||
|
||||
var fun = function(id_task, id_live_preview){
|
||||
request("/internal/progress", {"id_task": id_task, "id_live_preview": id_live_preview}, function(res){
|
||||
console.log(res)
|
||||
|
||||
if(res.completed){
|
||||
removeProgressBar()
|
||||
return
|
||||
@@ -184,15 +182,15 @@ function requestProgress(id_task, progressbarContainer, gallery, atEnd, onProgre
|
||||
|
||||
|
||||
if(res.live_preview){
|
||||
|
||||
var rect = gallery.getBoundingClientRect()
|
||||
if(rect.width){
|
||||
livePreview.style.width = rect.width + "px"
|
||||
livePreview.style.height = rect.height + "px"
|
||||
}
|
||||
|
||||
var img = new Image();
|
||||
img.onload = function() {
|
||||
var rect = gallery.getBoundingClientRect()
|
||||
if(rect.width){
|
||||
livePreview.style.width = rect.width + "px"
|
||||
livePreview.style.height = rect.height + "px"
|
||||
}
|
||||
|
||||
livePreview.innerHTML = ''
|
||||
livePreview.appendChild(img)
|
||||
if(livePreview.childElementCount > 2){
|
||||
livePreview.removeChild(livePreview.firstElementChild)
|
||||
@@ -208,7 +206,7 @@ function requestProgress(id_task, progressbarContainer, gallery, atEnd, onProgre
|
||||
|
||||
setTimeout(() => {
|
||||
fun(id_task, res.id_live_preview);
|
||||
}, 500)
|
||||
}, opts.live_preview_refresh_period || 500)
|
||||
}, function(){
|
||||
removeProgressBar()
|
||||
})
|
||||
|
||||
@@ -37,6 +37,9 @@ def quote(text):
|
||||
|
||||
|
||||
def image_from_url_text(filedata):
|
||||
if filedata is None:
|
||||
return None
|
||||
|
||||
if type(filedata) == list and len(filedata) > 0 and type(filedata[0]) == dict and filedata[0].get("is_file", False):
|
||||
filedata = filedata[0]
|
||||
|
||||
|
||||
@@ -561,6 +561,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi
|
||||
_loss_step = 0 #internal
|
||||
# size = len(ds.indexes)
|
||||
# loss_dict = defaultdict(lambda : deque(maxlen = 1024))
|
||||
loss_logging = deque(maxlen=len(ds) * 3) # this should be configurable parameter, this is 3 * epoch(dataset size)
|
||||
# losses = torch.zeros((size,))
|
||||
# previous_mean_losses = [0]
|
||||
# previous_mean_loss = 0
|
||||
@@ -610,7 +611,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi
|
||||
# go back until we reach gradient accumulation steps
|
||||
if (j + 1) % gradient_step != 0:
|
||||
continue
|
||||
|
||||
loss_logging.append(_loss_step)
|
||||
if clip_grad:
|
||||
clip_grad(weights, clip_grad_sched.learn_rate)
|
||||
|
||||
@@ -644,7 +645,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi
|
||||
if shared.opts.training_enable_tensorboard:
|
||||
epoch_num = hypernetwork.step // len(ds)
|
||||
epoch_step = hypernetwork.step - (epoch_num * len(ds)) + 1
|
||||
|
||||
mean_loss = sum(loss_logging) / len(loss_logging)
|
||||
textual_inversion.tensorboard_add(tensorboard_writer, loss=mean_loss, global_step=hypernetwork.step, step=epoch_step, learn_rate=scheduler.learn_rate, epoch_num=epoch_num)
|
||||
|
||||
textual_inversion.write_loss(log_directory, "hypernetwork_loss.csv", hypernetwork.step, steps_per_epoch, {
|
||||
@@ -688,9 +689,6 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi
|
||||
|
||||
processed = processing.process_images(p)
|
||||
image = processed.images[0] if len(processed.images) > 0 else None
|
||||
|
||||
if shared.opts.training_enable_tensorboard and shared.opts.training_tensorboard_save_images:
|
||||
textual_inversion.tensorboard_add_image(tensorboard_writer, f"Validation at epoch {epoch_num}", image, hypernetwork.step)
|
||||
|
||||
if unload:
|
||||
shared.sd_model.cond_stage_model.to(devices.cpu)
|
||||
@@ -701,7 +699,10 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi
|
||||
hypernetwork.train()
|
||||
if image is not None:
|
||||
shared.state.assign_current_image(image)
|
||||
|
||||
if shared.opts.training_enable_tensorboard and shared.opts.training_tensorboard_save_images:
|
||||
textual_inversion.tensorboard_add_image(tensorboard_writer,
|
||||
f"Validation at epoch {epoch_num}", image,
|
||||
hypernetwork.step)
|
||||
last_saved_image, last_text_info = images.save_image(image, images_dir, "", p.seed, p.prompt, shared.opts.samples_format, processed.infotexts[0], p=p, forced_filename=forced_filename, save_to_dirs=False)
|
||||
last_saved_image += f", prompt: {preview_text}"
|
||||
|
||||
|
||||
@@ -274,6 +274,7 @@ re_attention = re.compile(r"""
|
||||
:
|
||||
""", re.X)
|
||||
|
||||
re_break = re.compile(r"\s*\bBREAK\b\s*", re.S)
|
||||
|
||||
def parse_prompt_attention(text):
|
||||
"""
|
||||
@@ -339,7 +340,11 @@ def parse_prompt_attention(text):
|
||||
elif text == ']' and len(square_brackets) > 0:
|
||||
multiply_range(square_brackets.pop(), square_bracket_multiplier)
|
||||
else:
|
||||
res.append([text, 1.0])
|
||||
parts = re.split(re_break, text)
|
||||
for i, part in enumerate(parts):
|
||||
if i > 0:
|
||||
res.append(["BREAK", -1])
|
||||
res.append([part, 1.0])
|
||||
|
||||
for pos in round_brackets:
|
||||
multiply_range(pos, round_bracket_multiplier)
|
||||
|
||||
@@ -96,13 +96,18 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
|
||||
token_count = 0
|
||||
last_comma = -1
|
||||
|
||||
def next_chunk():
|
||||
"""puts current chunk into the list of results and produces the next one - empty"""
|
||||
def next_chunk(is_last=False):
|
||||
"""puts current chunk into the list of results and produces the next one - empty;
|
||||
if is_last is true, tokens <end-of-text> tokens at the end won't add to token_count"""
|
||||
nonlocal token_count
|
||||
nonlocal last_comma
|
||||
nonlocal chunk
|
||||
|
||||
token_count += len(chunk.tokens)
|
||||
if is_last:
|
||||
token_count += len(chunk.tokens)
|
||||
else:
|
||||
token_count += self.chunk_length
|
||||
|
||||
to_add = self.chunk_length - len(chunk.tokens)
|
||||
if to_add > 0:
|
||||
chunk.tokens += [self.id_end] * to_add
|
||||
@@ -116,6 +121,10 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
|
||||
chunk = PromptChunk()
|
||||
|
||||
for tokens, (text, weight) in zip(tokenized, parsed):
|
||||
if text == 'BREAK' and weight == -1:
|
||||
next_chunk()
|
||||
continue
|
||||
|
||||
position = 0
|
||||
while position < len(tokens):
|
||||
token = tokens[position]
|
||||
@@ -159,7 +168,7 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
|
||||
position += embedding_length_in_tokens
|
||||
|
||||
if len(chunk.tokens) > 0 or len(chunks) == 0:
|
||||
next_chunk()
|
||||
next_chunk(is_last=True)
|
||||
|
||||
return chunks, token_count
|
||||
|
||||
|
||||
+3
-1
@@ -117,6 +117,7 @@ restricted_opts = {
|
||||
}
|
||||
|
||||
ui_reorder_categories = [
|
||||
"masking",
|
||||
"sampler",
|
||||
"dimensions",
|
||||
"cfg",
|
||||
@@ -234,7 +235,7 @@ class State:
|
||||
if not parallel_processing_allowed:
|
||||
return
|
||||
|
||||
if self.sampling_step - self.current_image_sampling_step >= opts.show_progress_every_n_steps and opts.live_previews_enable:
|
||||
if self.sampling_step - self.current_image_sampling_step >= opts.show_progress_every_n_steps and opts.live_previews_enable and opts.show_progress_every_n_steps != -1:
|
||||
self.do_set_current_image()
|
||||
|
||||
def do_set_current_image(self):
|
||||
@@ -461,6 +462,7 @@ options_templates.update(options_section(('ui', "Live previews"), {
|
||||
"show_progress_every_n_steps": OptionInfo(10, "Show new live preview image every N sampling steps. Set to -1 to show after completion of batch.", gr.Slider, {"minimum": -1, "maximum": 32, "step": 1}),
|
||||
"show_progress_type": OptionInfo("Approx NN", "Image creation progress preview mode", gr.Radio, {"choices": ["Full", "Approx NN", "Approx cheap"]}),
|
||||
"live_preview_content": OptionInfo("Prompt", "Live preview subject", gr.Radio, {"choices": ["Combined", "Prompt", "Negative prompt"]}),
|
||||
"live_preview_refresh_period": OptionInfo(1000, "Progressbar/preview update period, in milliseconds")
|
||||
}))
|
||||
|
||||
options_templates.update(options_section(('sampler-params', "Sampler parameters"), {
|
||||
|
||||
@@ -2,7 +2,7 @@ import datetime
|
||||
import json
|
||||
import os
|
||||
|
||||
saved_params_shared = {"model_name", "model_hash", "initial_step", "num_of_dataset_images", "learn_rate", "batch_size", "clip_grad_mode", "clip_grad_value", "gradient_step", "data_root", "log_directory", "training_width", "training_height", "steps", "create_image_every", "template_file"}
|
||||
saved_params_shared = {"model_name", "model_hash", "initial_step", "num_of_dataset_images", "learn_rate", "batch_size", "clip_grad_mode", "clip_grad_value", "gradient_step", "data_root", "log_directory", "training_width", "training_height", "steps", "create_image_every", "template_file", "gradient_step", "latent_sampling_method"}
|
||||
saved_params_ti = {"embedding_name", "num_vectors_per_token", "save_embedding_every", "save_image_with_stored_embedding"}
|
||||
saved_params_hypernet = {"hypernetwork_name", "layer_structure", "activation_func", "weight_init", "add_layer_norm", "use_dropout", "save_hypernetwork_every"}
|
||||
saved_params_all = saved_params_shared | saved_params_ti | saved_params_hypernet
|
||||
|
||||
+30
-30
@@ -357,8 +357,8 @@ def create_toprow(is_img2img):
|
||||
|
||||
with gr.Column(scale=1):
|
||||
with gr.Row(elem_id=f"{id_part}_generate_box"):
|
||||
skip = gr.Button('Skip', elem_id=f"{id_part}_skip")
|
||||
interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt")
|
||||
skip = gr.Button('Skip', elem_id=f"{id_part}_skip")
|
||||
submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary')
|
||||
|
||||
skip.click(
|
||||
@@ -848,35 +848,6 @@ def create_ui():
|
||||
outputs=[],
|
||||
)
|
||||
|
||||
with FormGroup(elem_id="inpaint_controls", visible=False) as inpaint_controls:
|
||||
with FormRow():
|
||||
mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur")
|
||||
mask_alpha = gr.Slider(label="Mask transparency", visible=False, elem_id="img2img_mask_alpha")
|
||||
|
||||
with FormRow():
|
||||
inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode")
|
||||
|
||||
with FormRow():
|
||||
inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill")
|
||||
|
||||
with FormRow():
|
||||
with gr.Column():
|
||||
inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res")
|
||||
|
||||
with gr.Column(scale=4):
|
||||
inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding")
|
||||
|
||||
def select_img2img_tab(tab):
|
||||
return gr.update(visible=tab in [2, 3, 4]), gr.update(visible=tab == 3),
|
||||
|
||||
for i, elem in enumerate([tab_img2img, tab_sketch, tab_inpaint, tab_inpaint_color, tab_inpaint_upload, tab_batch]):
|
||||
elem.select(
|
||||
fn=lambda tab=i: select_img2img_tab(tab),
|
||||
inputs=[],
|
||||
outputs=[inpaint_controls, mask_alpha],
|
||||
)
|
||||
|
||||
|
||||
with FormRow():
|
||||
resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize")
|
||||
|
||||
@@ -918,6 +889,35 @@ def create_ui():
|
||||
with FormGroup(elem_id="img2img_script_container"):
|
||||
custom_inputs = modules.scripts.scripts_img2img.setup_ui()
|
||||
|
||||
elif category == "masking":
|
||||
with FormGroup(elem_id="inpaint_controls", visible=False) as inpaint_controls:
|
||||
with FormRow():
|
||||
mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur")
|
||||
mask_alpha = gr.Slider(label="Mask transparency", visible=False, elem_id="img2img_mask_alpha")
|
||||
|
||||
with FormRow():
|
||||
inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode")
|
||||
|
||||
with FormRow():
|
||||
inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill")
|
||||
|
||||
with FormRow():
|
||||
with gr.Column():
|
||||
inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res")
|
||||
|
||||
with gr.Column(scale=4):
|
||||
inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding")
|
||||
|
||||
def select_img2img_tab(tab):
|
||||
return gr.update(visible=tab in [2, 3, 4]), gr.update(visible=tab == 3),
|
||||
|
||||
for i, elem in enumerate([tab_img2img, tab_sketch, tab_inpaint, tab_inpaint_color, tab_inpaint_upload, tab_batch]):
|
||||
elem.select(
|
||||
fn=lambda tab=i: select_img2img_tab(tab),
|
||||
inputs=[],
|
||||
outputs=[inpaint_controls, mask_alpha],
|
||||
)
|
||||
|
||||
img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples)
|
||||
parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt)
|
||||
|
||||
|
||||
@@ -332,6 +332,7 @@ input[type="range"]{
|
||||
}
|
||||
|
||||
.livePreview img{
|
||||
position: absolute;
|
||||
object-fit: contain;
|
||||
width: 100%;
|
||||
height: 100%;
|
||||
@@ -473,13 +474,13 @@ input[type="range"]{
|
||||
}
|
||||
|
||||
#txt2img_interrupt, #img2img_interrupt{
|
||||
right: 0;
|
||||
border-radius: 0 0.5rem 0.5rem 0;
|
||||
}
|
||||
#txt2img_skip, #img2img_skip{
|
||||
left: 0;
|
||||
border-radius: 0.5rem 0 0 0.5rem;
|
||||
}
|
||||
#txt2img_skip, #img2img_skip{
|
||||
right: 0;
|
||||
border-radius: 0 0.5rem 0.5rem 0;
|
||||
}
|
||||
|
||||
.red {
|
||||
color: red;
|
||||
|
||||
Reference in New Issue
Block a user