diff options
Diffstat (limited to 'modules/hypernetworks')
| -rw-r--r-- | modules/hypernetworks/hypernetwork.py | 18 |
1 files changed, 12 insertions, 6 deletions
diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 503534e2..f6ef42d5 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -307,7 +307,7 @@ class Hypernetwork: def shorthash(self):
sha256 = hashes.sha256(self.filename, f'hypernet/{self.name}')
- return sha256[0:10]
+ return sha256[0:10] if sha256 else None
def list_hypernetworks(path):
@@ -380,8 +380,8 @@ def apply_single_hypernetwork(hypernetwork, context_k, context_v, layer=None): layer.hyper_k = hypernetwork_layers[0]
layer.hyper_v = hypernetwork_layers[1]
- context_k = hypernetwork_layers[0](context_k)
- context_v = hypernetwork_layers[1](context_v)
+ context_k = devices.cond_cast_unet(hypernetwork_layers[0](devices.cond_cast_float(context_k)))
+ context_v = devices.cond_cast_unet(hypernetwork_layers[1](devices.cond_cast_float(context_v)))
return context_k, context_v
@@ -496,7 +496,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, shared.reload_hypernetworks()
-def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_filename, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height):
+def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, use_weight, create_image_every, save_hypernetwork_every, template_filename, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height):
# images allows training previews to have infotext. Importing it at the top causes a circular import problem.
from modules import images
@@ -554,7 +554,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi pin_memory = shared.opts.pin_memory
- ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method, varsize=varsize)
+ ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method, varsize=varsize, use_weight=use_weight)
if shared.opts.save_training_settings_to_txt:
saved_params = dict(
@@ -640,13 +640,19 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi with devices.autocast():
x = batch.latent_sample.to(devices.device, non_blocking=pin_memory)
+ if use_weight:
+ w = batch.weight.to(devices.device, non_blocking=pin_memory)
if tag_drop_out != 0 or shuffle_tags:
shared.sd_model.cond_stage_model.to(devices.device)
c = shared.sd_model.cond_stage_model(batch.cond_text).to(devices.device, non_blocking=pin_memory)
shared.sd_model.cond_stage_model.to(devices.cpu)
else:
c = stack_conds(batch.cond).to(devices.device, non_blocking=pin_memory)
- loss = shared.sd_model(x, c)[0] / gradient_step
+ if use_weight:
+ loss = shared.sd_model.weighted_forward(x, c, w)[0] / gradient_step
+ del w
+ else:
+ loss = shared.sd_model.forward(x, c)[0] / gradient_step
del x
del c
|