Added anchors to regulate the lora

README.md

@@ -104,8 +104,18 @@ Just went in and out. It is much worse on smaller faces than shown here.
 
 <img src="https://raw.githubusercontent.com/ostris/ai-toolkit/main/assets/VAE_test1.jpg" width="768" height="auto"> 
 
+---
+
 ## TODO
-- [ ] Add proper regs on sliders
+- [X] Add proper regs on sliders
 - [ ] Add SDXL support (base model only for now)
 - [ ] Add plain erasing
-- [ ] Make Textual inversion network trainer (network that spits out TI embeddings)
+- [ ] Make Textual inversion network trainer (network that spits out TI embeddings)
+
+---
+
+## Change Log
+#### 2021-07-30
+Added "anchors" to the slider trainer. This allows you to set a prompt that will be used as a 
+regularizer. You can set the network multiplier to force spread consistency at high weights
+

config/examples/train_slider.example.yml

@@ -163,6 +163,25 @@ config:
             # to a lower number like 0.1 so they dont outweigh the primary target
             weight: 1.0
 
+        # anchors are prompts that wer try to hold on to while training the slider
+        # you want these to generate an image very similar to the target_class
+        # without directly overlapping it. For example, if you are training on a person smiling,
+        # you would use "a person with a face mask" as an anchor. It is a person, the image is the same
+        # regardless if they are smiling or not
+        anchors:
+          # only positive prompt for now
+          - prompt: "a woman"
+            neg_prompt: "animal"
+            # the multiplier applied to the LoRA when this is run.
+            # higher will give it more weight but also help keep the lora from collapsing
+            multiplier: 8.0
+          - prompt: "a man"
+            neg_prompt: "animal"
+            multiplier: 8.0
+          - prompt: "a person"
+            neg_prompt: "animal"
+            multiplier: 8.0
+
 # You can put any information you want here, and it will be saved in the model.
 # The below is an example, but you can put your grocery list in it if you want.
 # It is saved in the model so be aware of that. The software will include this

jobs/process/BaseSDTrainProcess.py

@@ -2,6 +2,7 @@ import time
 from collections import OrderedDict
 import os
 
+from leco.train_util import predict_noise
 from toolkit.kohya_model_util import load_vae
 from toolkit.lora_special import LoRASpecialNetwork
 from toolkit.optimizer import get_optimizer
@@ -59,11 +60,10 @@ class BaseSDTrainProcess(BaseTrainProcess):
         self.logging_config = LogingConfig(**self.get_conf('logging', {}))
         self.optimizer = None
         self.lr_scheduler = None
-        self.sd = None
+        self.sd: 'StableDiffusion' = None
 
         # added later
         self.network = None
-        self.scheduler = None
 
     def sample(self, step=None):
         sample_folder = os.path.join(self.save_root, 'samples')
@@ -118,7 +118,7 @@ class BaseSDTrainProcess(BaseTrainProcess):
                             'multiplier': self.network.multiplier,
                         })
 
-                for i in tqdm(range(len(self.sample_config.prompts)), desc=f"Generating Samples - step: {step}"):
+                for i in tqdm(range(len(self.sample_config.prompts)), desc=f"Generating Samples - step: {step}", leave=False):
                     raw_prompt = self.sample_config.prompts[i]
 
                     neg = self.sample_config.neg
@@ -267,6 +267,27 @@ class BaseSDTrainProcess(BaseTrainProcess):
         # return loss
         return 0.0
 
+    # ref: https://github.com/huggingface/diffusers/blob/0bab447670f47c28df60fbd2f6a0f833f75a16f5/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py#L746
+    def diffuse_some_steps(
+            self,
+            latents: torch.FloatTensor,
+            text_embeddings: torch.FloatTensor,
+            total_timesteps: int = 1000,
+            start_timesteps=0,
+            **kwargs,
+    ):
+
+        for timestep in tqdm(self.sd.noise_scheduler.timesteps[start_timesteps:total_timesteps], leave=False):
+            noise_pred = train_util.predict_noise(
+                self.sd.unet, self.sd.noise_scheduler, timestep, latents, text_embeddings, **kwargs
+            )
+
+            # compute the previous noisy sample x_t -> x_t-1
+            latents = self.sd.noise_scheduler.step(noise_pred, timestep, latents).prev_sample
+
+        # return latents_steps
+        return latents
+
     def run(self):
         super().run()
 
@@ -368,7 +389,21 @@ class BaseSDTrainProcess(BaseTrainProcess):
             # todo handle dataloader here maybe, not sure
 
             ### HOOK ###
-            loss = self.hook_train_loop()
+            loss_dict = self.hook_train_loop()
+
+            if self.train_config.optimizer.startswith('dadaptation'):
+                learning_rate = (
+                        optimizer.param_groups[0]["d"] *
+                        optimizer.param_groups[0]["lr"]
+                )
+            else:
+                learning_rate = optimizer.param_groups[0]['lr']
+
+            prog_bar_string = f"lr: {learning_rate:.1e}"
+            for key, value in loss_dict.items():
+                prog_bar_string += f" {key}: {value:.3e}"
+
+            self.progress_bar.set_postfix_str(prog_bar_string)
 
             # don't do on first step
             if self.step_num != self.start_step:
@@ -386,15 +421,8 @@ class BaseSDTrainProcess(BaseTrainProcess):
                 if self.logging_config.log_every and self.step_num % self.logging_config.log_every == 0:
                     # log to tensorboard
                     if self.writer is not None:
-                        # get avg loss
-                        self.writer.add_scalar(f"loss", loss, self.step_num)
-                        if self.train_config.optimizer.startswith('dadaptation'):
-                            learning_rate = (
-                                    optimizer.param_groups[0]["d"] *
-                                    optimizer.param_groups[0]["lr"]
-                            )
-                        else:
-                            learning_rate = optimizer.param_groups[0]['lr']
+                        for key, value in loss_dict.items():
+                            self.writer.add_scalar(f"{key}", value, self.step_num)
                         self.writer.add_scalar(f"lr", learning_rate, self.step_num)
                 self.progress_bar.refresh()
 

jobs/process/TrainSliderProcess.py

@@ -29,6 +29,7 @@ def flush():
     gc.collect()
 
 
+
 class EncodedPromptPair:
     def __init__(
             self,
@@ -53,6 +54,18 @@ class EncodedPromptPair:
         self.weight = weight
 
 
+class EncodedAnchor:
+    def __init__(
+            self,
+            prompt,
+            neg_prompt,
+            multiplier=1.0
+    ):
+        self.prompt = prompt
+        self.neg_prompt = neg_prompt
+        self.multiplier = multiplier
+
+
 class TrainSliderProcess(BaseSDTrainProcess):
     def __init__(self, process_id: int, job, config: OrderedDict):
         super().__init__(process_id, job, config)
@@ -61,9 +74,9 @@ class TrainSliderProcess(BaseSDTrainProcess):
         self.device = self.get_conf('device', self.job.device)
         self.device_torch = torch.device(self.device)
         self.slider_config = SliderConfig(**self.get_conf('slider', {}))
-
         self.prompt_cache = PromptEmbedsCache()
         self.prompt_pairs: list[EncodedPromptPair] = []
+        self.anchor_pairs: list[EncodedAnchor] = []
 
     def before_model_load(self):
         pass
@@ -146,16 +159,39 @@ class TrainSliderProcess(BaseSDTrainProcess):
                         ),
                     ]
 
+            # setup anchors
+            anchor_pairs = []
+            for anchor in self.slider_config.anchors:
+                # build the cache
+                for prompt in [
+                    anchor.prompt,
+                    anchor.neg_prompt  # empty neutral
+                ]:
+                    if cache[prompt] == None:
+                        cache[prompt] = train_util.encode_prompts(
+                            self.sd.tokenizer, self.sd.text_encoder, [prompt]
+                        )
+
+                anchor_pairs += [
+                    EncodedAnchor(
+                        prompt=cache[anchor.prompt],
+                        neg_prompt=cache[anchor.neg_prompt],
+                        multiplier=anchor.multiplier
+                    )
+                ]
+
         # move to cpu to save vram
         # We don't need text encoder anymore, but keep it on cpu for sampling
         self.sd.text_encoder.to("cpu")
         self.prompt_cache = cache
         self.prompt_pairs = prompt_pairs
+        self.anchor_pairs = anchor_pairs
         flush()
         # end hook_before_train_loop
 
     def hook_train_loop(self):
         dtype = get_torch_dtype(self.train_config.dtype)
+
         # get a random pair
         prompt_pair: EncodedPromptPair = self.prompt_pairs[
             torch.randint(0, len(self.prompt_pairs), (1,)).item()
@@ -202,10 +238,7 @@ class TrainSliderProcess(BaseSDTrainProcess):
 
             with self.network:
                 assert self.network.is_active
-                # A little denoised one is returned
-                denoised_latents = train_util.diffusion(
-                    unet,
-                    noise_scheduler,
+                denoised_latents = self.diffuse_some_steps(
                     latents,  # pass simple noise latents
                     train_util.concat_embeddings(
                         positive,  # unconditional
@@ -261,7 +294,46 @@ class TrainSliderProcess(BaseSDTrainProcess):
                 guidance_scale=1,
             ).to("cpu", dtype=torch.float32)
 
+        anchor_loss = None
+        if len(self.anchor_pairs) > 0:
+            # get a random anchor pair
+            anchor: EncodedAnchor = self.anchor_pairs[
+                torch.randint(0, len(self.anchor_pairs), (1,)).item()
+            ]
+            with torch.no_grad():
+                anchor_target_noise = train_util.predict_noise(
+                    unet,
+                    noise_scheduler,
+                    current_timestep,
+                    denoised_latents,
+                    train_util.concat_embeddings(
+                        anchor.prompt,
+                        anchor.neg_prompt,
+                        self.train_config.batch_size,
+                    ),
+                    guidance_scale=1,
+                ).to("cpu", dtype=torch.float32)
+            with self.network:
+                # anchor whatever weight  prompt pair is using
+                pos_nem_mult = 1.0 if prompt_pair.multiplier > 0 else -1.0
+                self.network.multiplier = anchor.multiplier * pos_nem_mult
+                anchor_pred_noise = train_util.predict_noise(
+                    unet,
+                    noise_scheduler,
+                    current_timestep,
+                    denoised_latents,
+                    train_util.concat_embeddings(
+                        anchor.prompt,
+                        anchor.neg_prompt,
+                        self.train_config.batch_size,
+                    ),
+                    guidance_scale=1,
+                ).to("cpu", dtype=torch.float32)
+
+                self.network.multiplier = prompt_pair.multiplier
+
         with self.network:
+            self.network.multiplier = prompt_pair.multiplier
             target_latents = train_util.predict_noise(
                 unet,
                 noise_scheduler,
@@ -281,7 +353,12 @@ class TrainSliderProcess(BaseSDTrainProcess):
         positive_latents.requires_grad = False
         neutral_latents.requires_grad = False
         unconditional_latents.requires_grad = False
-
+        if len(self.anchor_pairs) > 0:
+            anchor_target_noise.requires_grad = False
+            anchor_loss = loss_function(
+                anchor_target_noise,
+                anchor_pred_noise,
+            )
         erase = prompt_pair.action == ACTION_TYPES_SLIDER.ERASE_NEGATIVE
         guidance_scale = 1.0
 
@@ -299,16 +376,14 @@ class TrainSliderProcess(BaseSDTrainProcess):
             offset_neutral,
         ) * weight
 
-        loss_float = loss.item()
-        if self.train_config.optimizer.startswith('dadaptation'):
-            learning_rate = (
-                    optimizer.param_groups[0]["d"] *
-                    optimizer.param_groups[0]["lr"]
-            )
-        else:
-            learning_rate = optimizer.param_groups[0]['lr']
+        loss_slide = loss.item()
 
-        self.progress_bar.set_postfix_str(f"lr: {learning_rate:.1e} loss: {loss.item():.3e}")
+        if anchor_loss is not None:
+            loss += anchor_loss
+
+        loss_float = loss.item()
+
+        loss = loss.to(self.device_torch)
 
         loss.backward()
         optimizer.step()
@@ -326,5 +401,12 @@ class TrainSliderProcess(BaseSDTrainProcess):
         # reset network
         self.network.multiplier = 1.0
 
-        return loss_float
+        loss_dict = OrderedDict(
+            {'loss': loss_float},
+        )
+        if anchor_loss is not None:
+            loss_dict['sl_l'] = loss_slide
+            loss_dict['an_l'] = anchor_loss.item()
+
+        return loss_dict
         # end hook_train_loop

toolkit/config_modules.py

@@ -71,9 +71,19 @@ class SliderTargetConfig:
         self.weight: float = kwargs.get('weight', 1.0)
 
 
+class SliderConfigAnchors:
+    def __init__(self, **kwargs):
+        self.prompt = kwargs.get('prompt', '')
+        self.neg_prompt = kwargs.get('neg_prompt', '')
+        self.multiplier = kwargs.get('multiplier', 1.0)
+
+
 class SliderConfig:
     def __init__(self, **kwargs):
         targets = kwargs.get('targets', [])
         targets = [SliderTargetConfig(**target) for target in targets]
         self.targets: List[SliderTargetConfig] = targets
+        anchors = kwargs.get('anchors', [])
+        anchors = [SliderConfigAnchors(**anchor) for anchor in anchors]
+        self.anchors: List[SliderConfigAnchors] = anchors
         self.resolutions: List[List[int]] = kwargs.get('resolutions', [[512, 512]])