From a15b9b34f9a528844d4b5e752a23ea25edf216df Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Wed, 9 Oct 2024 19:45:15 +0530 Subject: [PATCH 01/12] log a warning when there are missing keys in the LoRA loading. --- src/diffusers/loaders/lora_pipeline.py | 34 +++++++++++++++++++++++++- src/diffusers/loaders/unet.py | 8 ++++++ 2 files changed, 41 insertions(+), 1 deletion(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index 2037bd787433..6bf3941f53a8 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -1367,6 +1367,14 @@ def load_lora_into_transformer( f" {unexpected_keys}. " ) + # Filter missing keys specific to the current adapter. + missing_keys = getattr(incompatible_keys, "missing_keys", None) + if missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model: " + f" {missing_keys}. " + ) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() @@ -1933,7 +1941,7 @@ def load_lora_into_transformer( incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name, **peft_kwargs) if incompatible_keys is not None: - # check only for unexpected keys + # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: logger.warning( @@ -1941,6 +1949,14 @@ def load_lora_into_transformer( f" {unexpected_keys}. " ) + # Filter missing keys specific to the current adapter. + missing_keys = getattr(incompatible_keys, "missing_keys", None) + if missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model: " + f" {missing_keys}. " + ) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() @@ -2288,6 +2304,14 @@ def load_lora_into_transformer(cls, state_dict, network_alphas, transformer, ada f" {unexpected_keys}. " ) + # Filter missing keys specific to the current adapter. + missing_keys = getattr(incompatible_keys, "missing_keys", None) + if missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model: " + f" {missing_keys}. " + ) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() @@ -2726,6 +2750,14 @@ def load_lora_into_transformer( f" {unexpected_keys}. " ) + # Filter missing keys specific to the current adapter. + missing_keys = getattr(incompatible_keys, "missing_keys", None) + if missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model: " + f" {missing_keys}. " + ) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() diff --git a/src/diffusers/loaders/unet.py b/src/diffusers/loaders/unet.py index eaac52df6202..57abfbba46d5 100644 --- a/src/diffusers/loaders/unet.py +++ b/src/diffusers/loaders/unet.py @@ -363,6 +363,14 @@ def _process_lora( f" {unexpected_keys}. " ) + # Filter missing keys specific to the current adapter. + missing_keys = getattr(incompatible_keys, "missing_keys", None) + if missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model: " + f" {missing_keys}. " + ) + return is_model_cpu_offload, is_sequential_cpu_offload @classmethod From 58da14ba9fbf926b29e1ddd49ad476c1efa66711 Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Wed, 9 Oct 2024 21:10:52 +0530 Subject: [PATCH 02/12] handle missing keys and unexpected keys better. --- src/diffusers/loaders/lora_pipeline.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index 6bf3941f53a8..1c1ae500bb98 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -1893,6 +1893,7 @@ def load_lora_into_transformer( state_dict = { k.replace(f"{cls.transformer_name}.", ""): v for k, v in state_dict.items() if k in transformer_keys } + print(f'{any("transformer_blocks." in k for k in state_dict)=}') if len(state_dict.keys()) > 0: # check with first key if is not in peft format @@ -1944,17 +1945,19 @@ def load_lora_into_transformer( # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: + lora_unexpected_keys = [k for k in unexpected_keys if "lora" in k and adapter_name in k] logger.warning( f"Loading adapter weights from state_dict led to unexpected keys not found in the model: " - f" {unexpected_keys}. " + f" {lora_unexpected_keys}. " ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: + lora_missing_keys = [k for k in missing_keys if "lora" in k and adapter_name in k] logger.warning( f"Loading adapter weights from state_dict led to missing keys in the model: " - f" {missing_keys}. " + f" {lora_missing_keys}. " ) # Offload back. From 83a2b36d5598f4626f9a74f9240d8893cddb11fd Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Wed, 9 Oct 2024 23:01:18 +0530 Subject: [PATCH 03/12] add tests --- src/diffusers/loaders/lora_pipeline.py | 49 +++++++++++++----------- src/diffusers/loaders/unet.py | 12 +++--- tests/lora/utils.py | 52 +++++++++++++++++++++++++- 3 files changed, 84 insertions(+), 29 deletions(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index 1c1ae500bb98..7c579d3a7649 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -1359,20 +1359,22 @@ def load_lora_into_transformer( incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name, **peft_kwargs) if incompatible_keys is not None: - # check only for unexpected keys + # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: + lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model: " - f" {unexpected_keys}. " + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: + lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model: " - f" {missing_keys}. " + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}" ) # Offload back. @@ -1893,7 +1895,6 @@ def load_lora_into_transformer( state_dict = { k.replace(f"{cls.transformer_name}.", ""): v for k, v in state_dict.items() if k in transformer_keys } - print(f'{any("transformer_blocks." in k for k in state_dict)=}') if len(state_dict.keys()) > 0: # check with first key if is not in peft format @@ -1945,19 +1946,19 @@ def load_lora_into_transformer( # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: - lora_unexpected_keys = [k for k in unexpected_keys if "lora" in k and adapter_name in k] + lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model: " - f" {lora_unexpected_keys}. " + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: - lora_missing_keys = [k for k in missing_keys if "lora" in k and adapter_name in k] + lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model: " - f" {lora_missing_keys}. " + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}" ) # Offload back. @@ -2299,20 +2300,22 @@ def load_lora_into_transformer(cls, state_dict, network_alphas, transformer, ada incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name) if incompatible_keys is not None: - # check only for unexpected keys + # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: + lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model: " - f" {unexpected_keys}. " + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: + lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model: " - f" {missing_keys}. " + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}." ) # Offload back. @@ -2745,20 +2748,22 @@ def load_lora_into_transformer( incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name, **peft_kwargs) if incompatible_keys is not None: - # check only for unexpected keys + # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: + lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model: " - f" {unexpected_keys}. " + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: + lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model: " - f" {missing_keys}. " + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}." ) # Offload back. diff --git a/src/diffusers/loaders/unet.py b/src/diffusers/loaders/unet.py index 57abfbba46d5..6156b761c6c5 100644 --- a/src/diffusers/loaders/unet.py +++ b/src/diffusers/loaders/unet.py @@ -355,20 +355,22 @@ def _process_lora( incompatible_keys = set_peft_model_state_dict(self, state_dict, adapter_name, **peft_kwargs) if incompatible_keys is not None: - # check only for unexpected keys + # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: + lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model: " - f" {unexpected_keys}. " + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: + lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model: " - f" {missing_keys}. " + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}." ) return is_model_cpu_offload, is_sequential_cpu_offload diff --git a/tests/lora/utils.py b/tests/lora/utils.py index 9c982e8de37f..389b5143aa54 100644 --- a/tests/lora/utils.py +++ b/tests/lora/utils.py @@ -27,8 +27,10 @@ LCMScheduler, UNet2DConditionModel, ) +from diffusers.utils import logging from diffusers.utils.import_utils import is_peft_available from diffusers.utils.testing_utils import ( + CaptureLogger, floats_tensor, require_peft_backend, require_peft_version_greater, @@ -219,10 +221,18 @@ def _get_modules_to_save(self, pipe, has_denoiser=False): modules_to_save = {} lora_loadable_modules = self.pipeline_class._lora_loadable_modules - if "text_encoder" in lora_loadable_modules and hasattr(pipe, "text_encoder"): + if ( + "text_encoder" in lora_loadable_modules + and hasattr(pipe, "text_encoder") + and getattr(pipe.text_encoder, "peft_config", None) is not None + ): modules_to_save["text_encoder"] = pipe.text_encoder - if "text_encoder_2" in lora_loadable_modules and hasattr(pipe, "text_encoder_2"): + if ( + "text_encoder_2" in lora_loadable_modules + and hasattr(pipe, "text_encoder_2") + and getattr(pipe.text_encoder_2, "peft_config", None) is not None + ): modules_to_save["text_encoder_2"] = pipe.text_encoder_2 if has_denoiser: @@ -1747,6 +1757,44 @@ def test_simple_inference_with_dora(self): "DoRA lora should change the output", ) + def test_missing_keys_warning(self): + scheduler_cls = self.scheduler_classes[0] + # Skip text encoder check for now as that is handled with `transformers`. + components, _, denoiser_lora_config = self.get_dummy_components(scheduler_cls) + pipe = self.pipeline_class(**components) + pipe = pipe.to(torch_device) + pipe.set_progress_bar_config(disable=None) + + denoiser = pipe.transformer if self.unet_kwargs is None else pipe.unet + denoiser.add_adapter(denoiser_lora_config) + self.assertTrue(check_if_lora_correctly_set(denoiser), "Lora not correctly set in denoiser.") + + with tempfile.TemporaryDirectory() as tmpdirname: + modules_to_save = self._get_modules_to_save(pipe, has_denoiser=True) + lora_state_dicts = self._get_lora_state_dicts(modules_to_save) + self.pipeline_class.save_lora_weights( + save_directory=tmpdirname, safe_serialization=False, **lora_state_dicts + ) + + self.assertTrue(os.path.isfile(os.path.join(tmpdirname, "pytorch_lora_weights.bin"))) + pipe.unload_lora_weights() + state_dict = torch.load(os.path.join(tmpdirname, "pytorch_lora_weights.bin"), weights_only=True) + + # To make things dynamic since we cannot settle with a single key for all the models where we + # offer PEFT support. + missing_key = [k for k in state_dict if "lora_A" in k][0] + del state_dict[missing_key] + + logger = logging.get_logger("diffusers.loaders.lora_pipeline") + logger.setLevel(logging.WARNING) + with CaptureLogger(logger) as cap_logger: + pipe.load_lora_weights(state_dict) + + # Since the missing key won't contain the adapter name ("default_0"). + # Also strip out the component prefix (such as "unet." from `missing_key`). + component = list({k.split(".")[0] for k in state_dict})[0] + self.assertTrue(missing_key.replace(f"{component}.", "") in str(cap_logger.out.replace("default_0.", ""))) + @unittest.skip("This is failing for now - need to investigate") def test_simple_inference_with_text_denoiser_lora_unfused_torch_compile(self): """ From e593d9eb8b5445d6f860761f461bbd9641847661 Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Wed, 9 Oct 2024 23:04:24 +0530 Subject: [PATCH 04/12] fix-copies. --- src/diffusers/loaders/lora_pipeline.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index 7c579d3a7649..c094f933d040 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -2763,7 +2763,7 @@ def load_lora_into_transformer( lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] logger.warning( f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}." + f" {', '.join(lora_missing_keys)}" ) # Offload back. From 0d47a1df5d257b84975f877848264eb3b7c8cad3 Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Wed, 9 Oct 2024 23:57:47 +0530 Subject: [PATCH 05/12] updates --- src/diffusers/loaders/lora_pipeline.py | 72 ++++++++++++++------------ src/diffusers/loaders/unet.py | 18 ++++--- 2 files changed, 50 insertions(+), 40 deletions(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index c094f933d040..630b4171106b 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -1363,19 +1363,21 @@ def load_lora_into_transformer( unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." - ) + if lora_unexpected_keys: + logger.warning( + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." + ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}" - ) + if lora_missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}" + ) # Offload back. if is_model_cpu_offload: @@ -1947,19 +1949,21 @@ def load_lora_into_transformer( unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." - ) + if lora_unexpected_keys: + logger.warning( + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." + ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}" - ) + if lora_missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}" + ) # Offload back. if is_model_cpu_offload: @@ -2304,19 +2308,21 @@ def load_lora_into_transformer(cls, state_dict, network_alphas, transformer, ada unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." - ) + if lora_unexpected_keys: + logger.warning( + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." + ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}." - ) + if lora_missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}." + ) # Offload back. if is_model_cpu_offload: @@ -2752,19 +2758,21 @@ def load_lora_into_transformer( unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." - ) + if lora_unexpected_keys: + logger.warning( + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." + ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}" - ) + if lora_missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}" + ) # Offload back. if is_model_cpu_offload: diff --git a/src/diffusers/loaders/unet.py b/src/diffusers/loaders/unet.py index 6156b761c6c5..64f4dd0ae6f6 100644 --- a/src/diffusers/loaders/unet.py +++ b/src/diffusers/loaders/unet.py @@ -359,19 +359,21 @@ def _process_lora( unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." - ) + if lora_unexpected_keys: + logger.warning( + f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f" {', '.join(lora_unexpected_keys)}." + ) # Filter missing keys specific to the current adapter. missing_keys = getattr(incompatible_keys, "missing_keys", None) if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] - logger.warning( - f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}." - ) + if lora_missing_keys: + logger.warning( + f"Loading adapter weights from state_dict led to missing keys in the model:" + f" {', '.join(lora_missing_keys)}." + ) return is_model_cpu_offload, is_sequential_cpu_offload From 33b3b9fd8c3a213013fd04e88010514f677ad1ad Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Thu, 10 Oct 2024 01:28:12 +0530 Subject: [PATCH 06/12] tests --- tests/lora/utils.py | 49 ++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 44 insertions(+), 5 deletions(-) diff --git a/tests/lora/utils.py b/tests/lora/utils.py index 389b5143aa54..3ff983773fb7 100644 --- a/tests/lora/utils.py +++ b/tests/lora/utils.py @@ -1775,9 +1775,8 @@ def test_missing_keys_warning(self): self.pipeline_class.save_lora_weights( save_directory=tmpdirname, safe_serialization=False, **lora_state_dicts ) - - self.assertTrue(os.path.isfile(os.path.join(tmpdirname, "pytorch_lora_weights.bin"))) pipe.unload_lora_weights() + self.assertTrue(os.path.isfile(os.path.join(tmpdirname, "pytorch_lora_weights.bin"))) state_dict = torch.load(os.path.join(tmpdirname, "pytorch_lora_weights.bin"), weights_only=True) # To make things dynamic since we cannot settle with a single key for all the models where we @@ -1785,15 +1784,55 @@ def test_missing_keys_warning(self): missing_key = [k for k in state_dict if "lora_A" in k][0] del state_dict[missing_key] - logger = logging.get_logger("diffusers.loaders.lora_pipeline") - logger.setLevel(logging.WARNING) + logger = ( + logging.get_logger("diffusers.loaders.unet") + if self.unet_kwargs is not None + else logging.get_logger("diffusers.loaders.lora_pipeline") + ) + logger.setLevel(30) with CaptureLogger(logger) as cap_logger: pipe.load_lora_weights(state_dict) # Since the missing key won't contain the adapter name ("default_0"). # Also strip out the component prefix (such as "unet." from `missing_key`). component = list({k.split(".")[0] for k in state_dict})[0] - self.assertTrue(missing_key.replace(f"{component}.", "") in str(cap_logger.out.replace("default_0.", ""))) + self.assertTrue(missing_key.replace(f"{component}.", "") in cap_logger.out.replace("default_0.", "")) + + def test_unexpected_keys_warning(self): + scheduler_cls = self.scheduler_classes[0] + # Skip text encoder check for now as that is handled with `transformers`. + components, _, denoiser_lora_config = self.get_dummy_components(scheduler_cls) + pipe = self.pipeline_class(**components) + pipe = pipe.to(torch_device) + pipe.set_progress_bar_config(disable=None) + + denoiser = pipe.transformer if self.unet_kwargs is None else pipe.unet + denoiser.add_adapter(denoiser_lora_config) + self.assertTrue(check_if_lora_correctly_set(denoiser), "Lora not correctly set in denoiser.") + + with tempfile.TemporaryDirectory() as tmpdirname: + modules_to_save = self._get_modules_to_save(pipe, has_denoiser=True) + lora_state_dicts = self._get_lora_state_dicts(modules_to_save) + self.pipeline_class.save_lora_weights( + save_directory=tmpdirname, safe_serialization=False, **lora_state_dicts + ) + pipe.unload_lora_weights() + self.assertTrue(os.path.isfile(os.path.join(tmpdirname, "pytorch_lora_weights.bin"))) + state_dict = torch.load(os.path.join(tmpdirname, "pytorch_lora_weights.bin"), weights_only=True) + + unexpected_key = [k for k in state_dict if "lora_A" in k][0] + ".diffusers_cat" + state_dict[unexpected_key] = torch.tensor(1.0, device=torch_device) + + logger = ( + logging.get_logger("diffusers.loaders.unet") + if self.unet_kwargs is not None + else logging.get_logger("diffusers.loaders.lora_pipeline") + ) + logger.setLevel(30) + with CaptureLogger(logger) as cap_logger: + pipe.load_lora_weights(state_dict) + + self.assertTrue(".diffusers_cat" in cap_logger.out) @unittest.skip("This is failing for now - need to investigate") def test_simple_inference_with_text_denoiser_lora_unfused_torch_compile(self): From 09989d87f3fa27214a65a9fa9d96a4329350d387 Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Thu, 10 Oct 2024 20:40:45 +0530 Subject: [PATCH 07/12] concat warning. --- src/diffusers/loaders/lora_pipeline.py | 46 +++++++++++++++++--------- src/diffusers/loaders/unet.py | 10 ++++-- 2 files changed, 38 insertions(+), 18 deletions(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index 630b4171106b..39414c122c26 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -1358,15 +1358,16 @@ def load_lora_into_transformer( inject_adapter_in_model(lora_config, transformer, adapter_name=adapter_name, **peft_kwargs) incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name, **peft_kwargs) + warn_msg = "" if incompatible_keys is not None: # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: - logger.warning( + warn_msg = ( f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." + f" {', '.join(lora_unexpected_keys)}. " ) # Filter missing keys specific to the current adapter. @@ -1374,11 +1375,14 @@ def load_lora_into_transformer( if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] if lora_missing_keys: - logger.warning( + warn_msg += ( f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}" + f" {', '.join(lora_missing_keys)}." ) + if warn_msg: + logger.warning(warn_msg) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() @@ -1944,15 +1948,16 @@ def load_lora_into_transformer( inject_adapter_in_model(lora_config, transformer, adapter_name=adapter_name, **peft_kwargs) incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name, **peft_kwargs) + warn_msg = "" if incompatible_keys is not None: # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: - logger.warning( + warn_msg = ( f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." + f" {', '.join(lora_unexpected_keys)}. " ) # Filter missing keys specific to the current adapter. @@ -1960,11 +1965,14 @@ def load_lora_into_transformer( if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] if lora_missing_keys: - logger.warning( + warn_msg += ( f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}" + f" {', '.join(lora_missing_keys)}." ) + if warn_msg: + logger.warning(warn_msg) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() @@ -2303,15 +2311,16 @@ def load_lora_into_transformer(cls, state_dict, network_alphas, transformer, ada inject_adapter_in_model(lora_config, transformer, adapter_name=adapter_name) incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name) + warn_msg = "" if incompatible_keys is not None: # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: - logger.warning( + warn_msg = ( f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." + f" {', '.join(lora_unexpected_keys)}. " ) # Filter missing keys specific to the current adapter. @@ -2319,11 +2328,14 @@ def load_lora_into_transformer(cls, state_dict, network_alphas, transformer, ada if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] if lora_missing_keys: - logger.warning( + warn_msg += ( f"Loading adapter weights from state_dict led to missing keys in the model:" f" {', '.join(lora_missing_keys)}." ) + if warn_msg: + logger.warning(warn_msg) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() @@ -2753,15 +2765,16 @@ def load_lora_into_transformer( inject_adapter_in_model(lora_config, transformer, adapter_name=adapter_name, **peft_kwargs) incompatible_keys = set_peft_model_state_dict(transformer, state_dict, adapter_name, **peft_kwargs) + warn_msg = "" if incompatible_keys is not None: # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: - logger.warning( + warn_msg = ( f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." + f" {', '.join(lora_unexpected_keys)}. " ) # Filter missing keys specific to the current adapter. @@ -2769,11 +2782,14 @@ def load_lora_into_transformer( if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] if lora_missing_keys: - logger.warning( + warn_msg += ( f"Loading adapter weights from state_dict led to missing keys in the model:" - f" {', '.join(lora_missing_keys)}" + f" {', '.join(lora_missing_keys)}." ) + if warn_msg: + logger.warning(warn_msg) + # Offload back. if is_model_cpu_offload: _pipeline.enable_model_cpu_offload() diff --git a/src/diffusers/loaders/unet.py b/src/diffusers/loaders/unet.py index 64f4dd0ae6f6..8f302bdad31b 100644 --- a/src/diffusers/loaders/unet.py +++ b/src/diffusers/loaders/unet.py @@ -354,15 +354,16 @@ def _process_lora( inject_adapter_in_model(lora_config, self, adapter_name=adapter_name, **peft_kwargs) incompatible_keys = set_peft_model_state_dict(self, state_dict, adapter_name, **peft_kwargs) + warn_msg = "" if incompatible_keys is not None: # Check only for unexpected keys. unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None) if unexpected_keys: lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: - logger.warning( + warn_msg = ( f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" - f" {', '.join(lora_unexpected_keys)}." + f" {', '.join(lora_unexpected_keys)}. " ) # Filter missing keys specific to the current adapter. @@ -370,11 +371,14 @@ def _process_lora( if missing_keys: lora_missing_keys = [k for k in missing_keys if "lora_" in k and adapter_name in k] if lora_missing_keys: - logger.warning( + warn_msg += ( f"Loading adapter weights from state_dict led to missing keys in the model:" f" {', '.join(lora_missing_keys)}." ) + if warn_msg: + logger.warning(warn_msg) + return is_model_cpu_offload, is_sequential_cpu_offload @classmethod From edade57e06c81d12c3f9de658195dcf5a19393d6 Mon Sep 17 00:00:00 2001 From: M Saqlain <118016760+saqlain2204@users.noreply.github.com> Date: Fri, 11 Oct 2024 19:17:31 +0530 Subject: [PATCH 08/12] Add Differential Diffusion to Kolors (#9423) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit * Added diff diff support for kolors img2img * Fized relative imports * Fized relative imports * Added diff diff support for Kolors * Fized import issues * Added map * Fized import issues * Fixed naming issues * Added diffdiff support for Kolors img2img pipeline * Removed example docstrings * Added map input * Updated latents Co-authored-by: Álvaro Somoza * Updated `original_with_noise` Co-authored-by: Álvaro Somoza * Improved code quality --------- Co-authored-by: Álvaro Somoza --- .../pipeline_kolors_differential_img2img.py | 1280 +++++++++++++++++ 1 file changed, 1280 insertions(+) create mode 100644 examples/community/pipeline_kolors_differential_img2img.py diff --git a/examples/community/pipeline_kolors_differential_img2img.py b/examples/community/pipeline_kolors_differential_img2img.py new file mode 100644 index 000000000000..e5570248d22b --- /dev/null +++ b/examples/community/pipeline_kolors_differential_img2img.py @@ -0,0 +1,1280 @@ +# Copyright 2024 Stability AI, Kwai-Kolors Team and The HuggingFace Team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +import inspect +from typing import Any, Callable, Dict, List, Optional, Tuple, Union + +import PIL.Image +import torch +from transformers import CLIPImageProcessor, CLIPVisionModelWithProjection + +from diffusers.callbacks import MultiPipelineCallbacks, PipelineCallback +from diffusers.image_processor import PipelineImageInput, VaeImageProcessor +from diffusers.loaders import IPAdapterMixin, StableDiffusionXLLoraLoaderMixin +from diffusers.models import AutoencoderKL, ImageProjection, UNet2DConditionModel +from diffusers.models.attention_processor import AttnProcessor2_0, FusedAttnProcessor2_0, XFormersAttnProcessor +from diffusers.pipelines.kolors.pipeline_output import KolorsPipelineOutput +from diffusers.pipelines.kolors.text_encoder import ChatGLMModel +from diffusers.pipelines.kolors.tokenizer import ChatGLMTokenizer +from diffusers.pipelines.pipeline_utils import DiffusionPipeline, StableDiffusionMixin +from diffusers.schedulers import KarrasDiffusionSchedulers +from diffusers.utils import is_torch_xla_available, logging, replace_example_docstring +from diffusers.utils.torch_utils import randn_tensor + + +if is_torch_xla_available(): + import torch_xla.core.xla_model as xm + + XLA_AVAILABLE = True +else: + XLA_AVAILABLE = False + + +logger = logging.get_logger(__name__) # pylint: disable=invalid-name + + +EXAMPLE_DOC_STRING = """ + Examples: + ```py + >>> import torch + >>> from diffusers import KolorsDifferentialImg2ImgPipeline + >>> from diffusers.utils import load_image + + >>> pipe = KolorsDifferentialImg2ImgPipeline.from_pretrained( + ... "Kwai-Kolors/Kolors-diffusers", variant="fp16", torch_dtype=torch.float16 + ... ) + >>> pipe = pipe.to("cuda") + >>> url = ( + ... "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/kolors/bunny_source.png" + ... ) + + + >>> init_image = load_image(url) + >>> prompt = "high quality image of a capybara wearing sunglasses. In the background of the image there are trees, poles, grass and other objects. At the bottom of the object there is the road., 8k, highly detailed." + >>> image = pipe(prompt, image=init_image).images[0] + ``` +""" + + +# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.retrieve_latents +def retrieve_latents( + encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample" +): + if hasattr(encoder_output, "latent_dist") and sample_mode == "sample": + return encoder_output.latent_dist.sample(generator) + elif hasattr(encoder_output, "latent_dist") and sample_mode == "argmax": + return encoder_output.latent_dist.mode() + elif hasattr(encoder_output, "latents"): + return encoder_output.latents + else: + raise AttributeError("Could not access latents of provided encoder_output") + + +# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps +def retrieve_timesteps( + scheduler, + num_inference_steps: Optional[int] = None, + device: Optional[Union[str, torch.device]] = None, + timesteps: Optional[List[int]] = None, + sigmas: Optional[List[float]] = None, + **kwargs, +): + """ + Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles + custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`. + + Args: + scheduler (`SchedulerMixin`): + The scheduler to get timesteps from. + num_inference_steps (`int`): + The number of diffusion steps used when generating samples with a pre-trained model. If used, `timesteps` + must be `None`. + device (`str` or `torch.device`, *optional*): + The device to which the timesteps should be moved to. If `None`, the timesteps are not moved. + timesteps (`List[int]`, *optional*): + Custom timesteps used to override the timestep spacing strategy of the scheduler. If `timesteps` is passed, + `num_inference_steps` and `sigmas` must be `None`. + sigmas (`List[float]`, *optional*): + Custom sigmas used to override the timestep spacing strategy of the scheduler. If `sigmas` is passed, + `num_inference_steps` and `timesteps` must be `None`. + + Returns: + `Tuple[torch.Tensor, int]`: A tuple where the first element is the timestep schedule from the scheduler and the + second element is the number of inference steps. + """ + if timesteps is not None and sigmas is not None: + raise ValueError("Only one of `timesteps` or `sigmas` can be passed. Please choose one to set custom values") + if timesteps is not None: + accepts_timesteps = "timesteps" in set(inspect.signature(scheduler.set_timesteps).parameters.keys()) + if not accepts_timesteps: + raise ValueError( + f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom" + f" timestep schedules. Please check whether you are using the correct scheduler." + ) + scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs) + timesteps = scheduler.timesteps + num_inference_steps = len(timesteps) + elif sigmas is not None: + accept_sigmas = "sigmas" in set(inspect.signature(scheduler.set_timesteps).parameters.keys()) + if not accept_sigmas: + raise ValueError( + f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom" + f" sigmas schedules. Please check whether you are using the correct scheduler." + ) + scheduler.set_timesteps(sigmas=sigmas, device=device, **kwargs) + timesteps = scheduler.timesteps + num_inference_steps = len(timesteps) + else: + scheduler.set_timesteps(num_inference_steps, device=device, **kwargs) + timesteps = scheduler.timesteps + return timesteps, num_inference_steps + + +class KolorsDifferentialImg2ImgPipeline( + DiffusionPipeline, StableDiffusionMixin, StableDiffusionXLLoraLoaderMixin, IPAdapterMixin +): + r""" + Pipeline for text-to-image generation using Kolors. + + This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods the + library implements for all the pipelines (such as downloading or saving, running on a particular device, etc.) + + The pipeline also inherits the following loading methods: + - [`~loaders.StableDiffusionXLLoraLoaderMixin.load_lora_weights`] for loading LoRA weights + - [`~loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights`] for saving LoRA weights + - [`~loaders.IPAdapterMixin.load_ip_adapter`] for loading IP Adapters + + Args: + vae ([`AutoencoderKL`]): + Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations. + text_encoder ([`ChatGLMModel`]): + Frozen text-encoder. Kolors uses [ChatGLM3-6B](https://huggingface.co/THUDM/chatglm3-6b). + tokenizer (`ChatGLMTokenizer`): + Tokenizer of class + [ChatGLMTokenizer](https://huggingface.co/THUDM/chatglm3-6b/blob/main/tokenization_chatglm.py). + unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents. + scheduler ([`SchedulerMixin`]): + A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of + [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`]. + force_zeros_for_empty_prompt (`bool`, *optional*, defaults to `"False"`): + Whether the negative prompt embeddings shall be forced to always be set to 0. Also see the config of + `Kwai-Kolors/Kolors-diffusers`. + """ + + model_cpu_offload_seq = "text_encoder->image_encoder-unet->vae" + _optional_components = [ + "image_encoder", + "feature_extractor", + ] + _callback_tensor_inputs = [ + "latents", + "prompt_embeds", + "negative_prompt_embeds", + "add_text_embeds", + "add_time_ids", + "negative_pooled_prompt_embeds", + "negative_add_time_ids", + ] + + def __init__( + self, + vae: AutoencoderKL, + text_encoder: ChatGLMModel, + tokenizer: ChatGLMTokenizer, + unet: UNet2DConditionModel, + scheduler: KarrasDiffusionSchedulers, + image_encoder: CLIPVisionModelWithProjection = None, + feature_extractor: CLIPImageProcessor = None, + force_zeros_for_empty_prompt: bool = False, + ): + super().__init__() + + self.register_modules( + vae=vae, + text_encoder=text_encoder, + tokenizer=tokenizer, + unet=unet, + scheduler=scheduler, + image_encoder=image_encoder, + feature_extractor=feature_extractor, + ) + self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt) + self.vae_scale_factor = ( + 2 ** (len(self.vae.config.block_out_channels) - 1) if hasattr(self, "vae") and self.vae is not None else 8 + ) + self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) + + self.mask_processor = VaeImageProcessor( + vae_scale_factor=self.vae_scale_factor, do_normalize=False, do_convert_grayscale=True + ) + + self.default_sample_size = self.unet.config.sample_size + + # Copied from diffusers.pipelines.kolors.pipeline_kolors.KolorsPipeline.encode_prompt + def encode_prompt( + self, + prompt, + device: Optional[torch.device] = None, + num_images_per_prompt: int = 1, + do_classifier_free_guidance: bool = True, + negative_prompt=None, + prompt_embeds: Optional[torch.FloatTensor] = None, + pooled_prompt_embeds: Optional[torch.Tensor] = None, + negative_prompt_embeds: Optional[torch.FloatTensor] = None, + negative_pooled_prompt_embeds: Optional[torch.Tensor] = None, + max_sequence_length: int = 256, + ): + r""" + Encodes the prompt into text encoder hidden states. + + Args: + prompt (`str` or `List[str]`, *optional*): + prompt to be encoded + device: (`torch.device`): + torch device + num_images_per_prompt (`int`): + number of images that should be generated per prompt + do_classifier_free_guidance (`bool`): + whether to use classifier free guidance or not + negative_prompt (`str` or `List[str]`, *optional*): + The prompt or prompts not to guide the image generation. If not defined, one has to pass + `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is + less than `1`). + prompt_embeds (`torch.FloatTensor`, *optional*): + Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not + provided, text embeddings will be generated from `prompt` input argument. + pooled_prompt_embeds (`torch.Tensor`, *optional*): + Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. + If not provided, pooled text embeddings will be generated from `prompt` input argument. + negative_prompt_embeds (`torch.FloatTensor`, *optional*): + Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt + weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input + argument. + negative_pooled_prompt_embeds (`torch.Tensor`, *optional*): + Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt + weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt` + input argument. + max_sequence_length (`int` defaults to 256): Maximum sequence length to use with the `prompt`. + """ + # from IPython import embed; embed(); exit() + device = device or self._execution_device + + if prompt is not None and isinstance(prompt, str): + batch_size = 1 + elif prompt is not None and isinstance(prompt, list): + batch_size = len(prompt) + else: + batch_size = prompt_embeds.shape[0] + + # Define tokenizers and text encoders + tokenizers = [self.tokenizer] + text_encoders = [self.text_encoder] + + if prompt_embeds is None: + prompt_embeds_list = [] + for tokenizer, text_encoder in zip(tokenizers, text_encoders): + text_inputs = tokenizer( + prompt, + padding="max_length", + max_length=max_sequence_length, + truncation=True, + return_tensors="pt", + ).to(device) + output = text_encoder( + input_ids=text_inputs["input_ids"], + attention_mask=text_inputs["attention_mask"], + position_ids=text_inputs["position_ids"], + output_hidden_states=True, + ) + + # [max_sequence_length, batch, hidden_size] -> [batch, max_sequence_length, hidden_size] + # clone to have a contiguous tensor + prompt_embeds = output.hidden_states[-2].permute(1, 0, 2).clone() + # [max_sequence_length, batch, hidden_size] -> [batch, hidden_size] + pooled_prompt_embeds = output.hidden_states[-1][-1, :, :].clone() + bs_embed, seq_len, _ = prompt_embeds.shape + prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1) + prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1) + + prompt_embeds_list.append(prompt_embeds) + + prompt_embeds = prompt_embeds_list[0] + + # get unconditional embeddings for classifier free guidance + zero_out_negative_prompt = negative_prompt is None and self.config.force_zeros_for_empty_prompt + + if do_classifier_free_guidance and negative_prompt_embeds is None and zero_out_negative_prompt: + negative_prompt_embeds = torch.zeros_like(prompt_embeds) + elif do_classifier_free_guidance and negative_prompt_embeds is None: + uncond_tokens: List[str] + if negative_prompt is None: + uncond_tokens = [""] * batch_size + elif prompt is not None and type(prompt) is not type(negative_prompt): + raise TypeError( + f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !=" + f" {type(prompt)}." + ) + elif isinstance(negative_prompt, str): + uncond_tokens = [negative_prompt] + elif batch_size != len(negative_prompt): + raise ValueError( + f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:" + f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches" + " the batch size of `prompt`." + ) + else: + uncond_tokens = negative_prompt + + negative_prompt_embeds_list = [] + + for tokenizer, text_encoder in zip(tokenizers, text_encoders): + uncond_input = tokenizer( + uncond_tokens, + padding="max_length", + max_length=max_sequence_length, + truncation=True, + return_tensors="pt", + ).to(device) + output = text_encoder( + input_ids=uncond_input["input_ids"], + attention_mask=uncond_input["attention_mask"], + position_ids=uncond_input["position_ids"], + output_hidden_states=True, + ) + + # [max_sequence_length, batch, hidden_size] -> [batch, max_sequence_length, hidden_size] + # clone to have a contiguous tensor + negative_prompt_embeds = output.hidden_states[-2].permute(1, 0, 2).clone() + # [max_sequence_length, batch, hidden_size] -> [batch, hidden_size] + negative_pooled_prompt_embeds = output.hidden_states[-1][-1, :, :].clone() + + if do_classifier_free_guidance: + # duplicate unconditional embeddings for each generation per prompt, using mps friendly method + seq_len = negative_prompt_embeds.shape[1] + + negative_prompt_embeds = negative_prompt_embeds.to(dtype=text_encoder.dtype, device=device) + + negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1) + negative_prompt_embeds = negative_prompt_embeds.view( + batch_size * num_images_per_prompt, seq_len, -1 + ) + + negative_prompt_embeds_list.append(negative_prompt_embeds) + + negative_prompt_embeds = negative_prompt_embeds_list[0] + + bs_embed = pooled_prompt_embeds.shape[0] + pooled_prompt_embeds = pooled_prompt_embeds.repeat(1, num_images_per_prompt).view( + bs_embed * num_images_per_prompt, -1 + ) + + if do_classifier_free_guidance: + negative_pooled_prompt_embeds = negative_pooled_prompt_embeds.repeat(1, num_images_per_prompt).view( + bs_embed * num_images_per_prompt, -1 + ) + + return prompt_embeds, negative_prompt_embeds, pooled_prompt_embeds, negative_pooled_prompt_embeds + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_image + def encode_image(self, image, device, num_images_per_prompt, output_hidden_states=None): + dtype = next(self.image_encoder.parameters()).dtype + + if not isinstance(image, torch.Tensor): + image = self.feature_extractor(image, return_tensors="pt").pixel_values + + image = image.to(device=device, dtype=dtype) + if output_hidden_states: + image_enc_hidden_states = self.image_encoder(image, output_hidden_states=True).hidden_states[-2] + image_enc_hidden_states = image_enc_hidden_states.repeat_interleave(num_images_per_prompt, dim=0) + uncond_image_enc_hidden_states = self.image_encoder( + torch.zeros_like(image), output_hidden_states=True + ).hidden_states[-2] + uncond_image_enc_hidden_states = uncond_image_enc_hidden_states.repeat_interleave( + num_images_per_prompt, dim=0 + ) + return image_enc_hidden_states, uncond_image_enc_hidden_states + else: + image_embeds = self.image_encoder(image).image_embeds + image_embeds = image_embeds.repeat_interleave(num_images_per_prompt, dim=0) + uncond_image_embeds = torch.zeros_like(image_embeds) + + return image_embeds, uncond_image_embeds + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_ip_adapter_image_embeds + def prepare_ip_adapter_image_embeds( + self, ip_adapter_image, ip_adapter_image_embeds, device, num_images_per_prompt, do_classifier_free_guidance + ): + image_embeds = [] + if do_classifier_free_guidance: + negative_image_embeds = [] + if ip_adapter_image_embeds is None: + if not isinstance(ip_adapter_image, list): + ip_adapter_image = [ip_adapter_image] + + if len(ip_adapter_image) != len(self.unet.encoder_hid_proj.image_projection_layers): + raise ValueError( + f"`ip_adapter_image` must have same length as the number of IP Adapters. Got {len(ip_adapter_image)} images and {len(self.unet.encoder_hid_proj.image_projection_layers)} IP Adapters." + ) + + for single_ip_adapter_image, image_proj_layer in zip( + ip_adapter_image, self.unet.encoder_hid_proj.image_projection_layers + ): + output_hidden_state = not isinstance(image_proj_layer, ImageProjection) + single_image_embeds, single_negative_image_embeds = self.encode_image( + single_ip_adapter_image, device, 1, output_hidden_state + ) + + image_embeds.append(single_image_embeds[None, :]) + if do_classifier_free_guidance: + negative_image_embeds.append(single_negative_image_embeds[None, :]) + else: + for single_image_embeds in ip_adapter_image_embeds: + if do_classifier_free_guidance: + single_negative_image_embeds, single_image_embeds = single_image_embeds.chunk(2) + negative_image_embeds.append(single_negative_image_embeds) + image_embeds.append(single_image_embeds) + + ip_adapter_image_embeds = [] + for i, single_image_embeds in enumerate(image_embeds): + single_image_embeds = torch.cat([single_image_embeds] * num_images_per_prompt, dim=0) + if do_classifier_free_guidance: + single_negative_image_embeds = torch.cat([negative_image_embeds[i]] * num_images_per_prompt, dim=0) + single_image_embeds = torch.cat([single_negative_image_embeds, single_image_embeds], dim=0) + + single_image_embeds = single_image_embeds.to(device=device) + ip_adapter_image_embeds.append(single_image_embeds) + + return ip_adapter_image_embeds + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs + def prepare_extra_step_kwargs(self, generator, eta): + # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature + # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers. + # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502 + # and should be between [0, 1] + + accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys()) + extra_step_kwargs = {} + if accepts_eta: + extra_step_kwargs["eta"] = eta + + # check if the scheduler accepts generator + accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys()) + if accepts_generator: + extra_step_kwargs["generator"] = generator + return extra_step_kwargs + + def check_inputs( + self, + prompt, + strength, + num_inference_steps, + height, + width, + negative_prompt=None, + prompt_embeds=None, + pooled_prompt_embeds=None, + negative_prompt_embeds=None, + negative_pooled_prompt_embeds=None, + ip_adapter_image=None, + ip_adapter_image_embeds=None, + callback_on_step_end_tensor_inputs=None, + max_sequence_length=None, + ): + if strength < 0 or strength > 1: + raise ValueError(f"The value of strength should in [0.0, 1.0] but is {strength}") + + if not isinstance(num_inference_steps, int) or num_inference_steps <= 0: + raise ValueError( + f"`num_inference_steps` has to be a positive integer but is {num_inference_steps} of type" + f" {type(num_inference_steps)}." + ) + + if height % 8 != 0 or width % 8 != 0: + raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.") + + if callback_on_step_end_tensor_inputs is not None and not all( + k in self._callback_tensor_inputs for k in callback_on_step_end_tensor_inputs + ): + raise ValueError( + f"`callback_on_step_end_tensor_inputs` has to be in {self._callback_tensor_inputs}, but found {[k for k in callback_on_step_end_tensor_inputs if k not in self._callback_tensor_inputs]}" + ) + + if prompt is not None and prompt_embeds is not None: + raise ValueError( + f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to" + " only forward one of the two." + ) + elif prompt is None and prompt_embeds is None: + raise ValueError( + "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined." + ) + elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)): + raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}") + + if negative_prompt is not None and negative_prompt_embeds is not None: + raise ValueError( + f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:" + f" {negative_prompt_embeds}. Please make sure to only forward one of the two." + ) + + if prompt_embeds is not None and negative_prompt_embeds is not None: + if prompt_embeds.shape != negative_prompt_embeds.shape: + raise ValueError( + "`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but" + f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`" + f" {negative_prompt_embeds.shape}." + ) + + if prompt_embeds is not None and pooled_prompt_embeds is None: + raise ValueError( + "If `prompt_embeds` are provided, `pooled_prompt_embeds` also have to be passed. Make sure to generate `pooled_prompt_embeds` from the same text encoder that was used to generate `prompt_embeds`." + ) + + if negative_prompt_embeds is not None and negative_pooled_prompt_embeds is None: + raise ValueError( + "If `negative_prompt_embeds` are provided, `negative_pooled_prompt_embeds` also have to be passed. Make sure to generate `negative_pooled_prompt_embeds` from the same text encoder that was used to generate `negative_prompt_embeds`." + ) + + if ip_adapter_image is not None and ip_adapter_image_embeds is not None: + raise ValueError( + "Provide either `ip_adapter_image` or `ip_adapter_image_embeds`. Cannot leave both `ip_adapter_image` and `ip_adapter_image_embeds` defined." + ) + + if ip_adapter_image_embeds is not None: + if not isinstance(ip_adapter_image_embeds, list): + raise ValueError( + f"`ip_adapter_image_embeds` has to be of type `list` but is {type(ip_adapter_image_embeds)}" + ) + elif ip_adapter_image_embeds[0].ndim not in [3, 4]: + raise ValueError( + f"`ip_adapter_image_embeds` has to be a list of 3D or 4D tensors but is {ip_adapter_image_embeds[0].ndim}D" + ) + + if max_sequence_length is not None and max_sequence_length > 256: + raise ValueError(f"`max_sequence_length` cannot be greater than 256 but is {max_sequence_length}") + + # Copied from diffusers.pipelines.stable_diffusion_xl.pipeline_stable_diffusion_xl_img2img.StableDiffusionXLImg2ImgPipeline.get_timesteps + def get_timesteps(self, num_inference_steps, strength, device, denoising_start=None): + # get the original timestep using init_timestep + if denoising_start is None: + init_timestep = min(int(num_inference_steps * strength), num_inference_steps) + t_start = max(num_inference_steps - init_timestep, 0) + else: + t_start = 0 + + timesteps = self.scheduler.timesteps[t_start * self.scheduler.order :] + + # Strength is irrelevant if we directly request a timestep to start at; + # that is, strength is determined by the denoising_start instead. + if denoising_start is not None: + discrete_timestep_cutoff = int( + round( + self.scheduler.config.num_train_timesteps + - (denoising_start * self.scheduler.config.num_train_timesteps) + ) + ) + + num_inference_steps = (timesteps < discrete_timestep_cutoff).sum().item() + if self.scheduler.order == 2 and num_inference_steps % 2 == 0: + # if the scheduler is a 2nd order scheduler we might have to do +1 + # because `num_inference_steps` might be even given that every timestep + # (except the highest one) is duplicated. If `num_inference_steps` is even it would + # mean that we cut the timesteps in the middle of the denoising step + # (between 1st and 2nd derivative) which leads to incorrect results. By adding 1 + # we ensure that the denoising process always ends after the 2nd derivate step of the scheduler + num_inference_steps = num_inference_steps + 1 + + # because t_n+1 >= t_n, we slice the timesteps starting from the end + timesteps = timesteps[-num_inference_steps:] + return timesteps, num_inference_steps + + return timesteps, num_inference_steps - t_start + + # Copied from diffusers.pipelines.stable_diffusion_xl.pipeline_stable_diffusion_xl_img2img.StableDiffusionXLImg2ImgPipeline.prepare_latents + def prepare_latents( + self, image, timestep, batch_size, num_images_per_prompt, dtype, device, generator=None, add_noise=True + ): + if not isinstance(image, (torch.Tensor, PIL.Image.Image, list)): + raise ValueError( + f"`image` has to be of type `torch.Tensor`, `PIL.Image.Image` or list but is {type(image)}" + ) + + latents_mean = latents_std = None + if hasattr(self.vae.config, "latents_mean") and self.vae.config.latents_mean is not None: + latents_mean = torch.tensor(self.vae.config.latents_mean).view(1, 4, 1, 1) + if hasattr(self.vae.config, "latents_std") and self.vae.config.latents_std is not None: + latents_std = torch.tensor(self.vae.config.latents_std).view(1, 4, 1, 1) + + # Offload text encoder if `enable_model_cpu_offload` was enabled + if hasattr(self, "final_offload_hook") and self.final_offload_hook is not None: + self.text_encoder_2.to("cpu") + torch.cuda.empty_cache() + + image = image.to(device=device, dtype=dtype) + + batch_size = batch_size * num_images_per_prompt + + if image.shape[1] == 4: + init_latents = image + + else: + # make sure the VAE is in float32 mode, as it overflows in float16 + if self.vae.config.force_upcast: + image = image.float() + self.vae.to(dtype=torch.float32) + + if isinstance(generator, list) and len(generator) != batch_size: + raise ValueError( + f"You have passed a list of generators of length {len(generator)}, but requested an effective batch" + f" size of {batch_size}. Make sure the batch size matches the length of the generators." + ) + + elif isinstance(generator, list): + if image.shape[0] < batch_size and batch_size % image.shape[0] == 0: + image = torch.cat([image] * (batch_size // image.shape[0]), dim=0) + elif image.shape[0] < batch_size and batch_size % image.shape[0] != 0: + raise ValueError( + f"Cannot duplicate `image` of batch size {image.shape[0]} to effective batch_size {batch_size} " + ) + + init_latents = [ + retrieve_latents(self.vae.encode(image[i : i + 1]), generator=generator[i]) + for i in range(batch_size) + ] + init_latents = torch.cat(init_latents, dim=0) + else: + init_latents = retrieve_latents(self.vae.encode(image), generator=generator) + + if self.vae.config.force_upcast: + self.vae.to(dtype) + + init_latents = init_latents.to(dtype) + if latents_mean is not None and latents_std is not None: + latents_mean = latents_mean.to(device=device, dtype=dtype) + latents_std = latents_std.to(device=device, dtype=dtype) + init_latents = (init_latents - latents_mean) * self.vae.config.scaling_factor / latents_std + else: + init_latents = self.vae.config.scaling_factor * init_latents + + if batch_size > init_latents.shape[0] and batch_size % init_latents.shape[0] == 0: + # expand init_latents for batch_size + additional_image_per_prompt = batch_size // init_latents.shape[0] + init_latents = torch.cat([init_latents] * additional_image_per_prompt, dim=0) + elif batch_size > init_latents.shape[0] and batch_size % init_latents.shape[0] != 0: + raise ValueError( + f"Cannot duplicate `image` of batch size {init_latents.shape[0]} to {batch_size} text prompts." + ) + else: + init_latents = torch.cat([init_latents], dim=0) + + if add_noise: + shape = init_latents.shape + noise = randn_tensor(shape, generator=generator, device=device, dtype=dtype) + # get latents + init_latents = self.scheduler.add_noise(init_latents, noise, timestep) + + latents = init_latents + + return latents + + # Copied from diffusers.pipelines.stable_diffusion_xl.pipeline_stable_diffusion_xl.StableDiffusionXLPipeline._get_add_time_ids + def _get_add_time_ids( + self, original_size, crops_coords_top_left, target_size, dtype, text_encoder_projection_dim=None + ): + add_time_ids = list(original_size + crops_coords_top_left + target_size) + + passed_add_embed_dim = ( + self.unet.config.addition_time_embed_dim * len(add_time_ids) + text_encoder_projection_dim + ) + expected_add_embed_dim = self.unet.add_embedding.linear_1.in_features + + if expected_add_embed_dim != passed_add_embed_dim: + raise ValueError( + f"Model expects an added time embedding vector of length {expected_add_embed_dim}, but a vector of {passed_add_embed_dim} was created. The model has an incorrect config. Please check `unet.config.time_embedding_type` and `text_encoder_2.config.projection_dim`." + ) + + add_time_ids = torch.tensor([add_time_ids], dtype=dtype) + return add_time_ids + + # Copied from diffusers.pipelines.stable_diffusion_xl.pipeline_stable_diffusion_xl.StableDiffusionXLPipeline.upcast_vae + def upcast_vae(self): + dtype = self.vae.dtype + self.vae.to(dtype=torch.float32) + use_torch_2_0_or_xformers = isinstance( + self.vae.decoder.mid_block.attentions[0].processor, + ( + AttnProcessor2_0, + XFormersAttnProcessor, + FusedAttnProcessor2_0, + ), + ) + # if xformers or torch_2_0 is used attention block does not need + # to be in float32 which can save lots of memory + if use_torch_2_0_or_xformers: + self.vae.post_quant_conv.to(dtype) + self.vae.decoder.conv_in.to(dtype) + self.vae.decoder.mid_block.to(dtype) + + # Copied from diffusers.pipelines.latent_consistency_models.pipeline_latent_consistency_text2img.LatentConsistencyModelPipeline.get_guidance_scale_embedding + def get_guidance_scale_embedding( + self, w: torch.Tensor, embedding_dim: int = 512, dtype: torch.dtype = torch.float32 + ) -> torch.Tensor: + """ + See https://github.com/google-research/vdm/blob/dc27b98a554f65cdc654b800da5aa1846545d41b/model_vdm.py#L298 + + Args: + w (`torch.Tensor`): + Generate embedding vectors with a specified guidance scale to subsequently enrich timestep embeddings. + embedding_dim (`int`, *optional*, defaults to 512): + Dimension of the embeddings to generate. + dtype (`torch.dtype`, *optional*, defaults to `torch.float32`): + Data type of the generated embeddings. + + Returns: + `torch.Tensor`: Embedding vectors with shape `(len(w), embedding_dim)`. + """ + assert len(w.shape) == 1 + w = w * 1000.0 + + half_dim = embedding_dim // 2 + emb = torch.log(torch.tensor(10000.0)) / (half_dim - 1) + emb = torch.exp(torch.arange(half_dim, dtype=dtype) * -emb) + emb = w.to(dtype)[:, None] * emb[None, :] + emb = torch.cat([torch.sin(emb), torch.cos(emb)], dim=1) + if embedding_dim % 2 == 1: # zero pad + emb = torch.nn.functional.pad(emb, (0, 1)) + assert emb.shape == (w.shape[0], embedding_dim) + return emb + + @property + def guidance_scale(self): + return self._guidance_scale + + # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2) + # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1` + # corresponds to doing no classifier free guidance. + @property + def do_classifier_free_guidance(self): + return self._guidance_scale > 1 and self.unet.config.time_cond_proj_dim is None + + @property + def cross_attention_kwargs(self): + return self._cross_attention_kwargs + + @property + def denoising_start(self): + return self._denoising_start + + @property + def denoising_end(self): + return self._denoising_end + + @property + def num_timesteps(self): + return self._num_timesteps + + @property + def interrupt(self): + return self._interrupt + + @torch.no_grad() + @replace_example_docstring(EXAMPLE_DOC_STRING) + def __call__( + self, + prompt: Union[str, List[str]] = None, + image: PipelineImageInput = None, + strength: float = 0.3, + height: Optional[int] = None, + width: Optional[int] = None, + num_inference_steps: int = 50, + timesteps: List[int] = None, + sigmas: List[float] = None, + denoising_start: Optional[float] = None, + denoising_end: Optional[float] = None, + guidance_scale: float = 5.0, + negative_prompt: Optional[Union[str, List[str]]] = None, + num_images_per_prompt: Optional[int] = 1, + eta: float = 0.0, + generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None, + latents: Optional[torch.Tensor] = None, + prompt_embeds: Optional[torch.Tensor] = None, + pooled_prompt_embeds: Optional[torch.Tensor] = None, + negative_prompt_embeds: Optional[torch.Tensor] = None, + negative_pooled_prompt_embeds: Optional[torch.Tensor] = None, + ip_adapter_image: Optional[PipelineImageInput] = None, + ip_adapter_image_embeds: Optional[List[torch.Tensor]] = None, + output_type: Optional[str] = "pil", + return_dict: bool = True, + cross_attention_kwargs: Optional[Dict[str, Any]] = None, + original_size: Optional[Tuple[int, int]] = None, + crops_coords_top_left: Tuple[int, int] = (0, 0), + target_size: Optional[Tuple[int, int]] = None, + negative_original_size: Optional[Tuple[int, int]] = None, + negative_crops_coords_top_left: Tuple[int, int] = (0, 0), + negative_target_size: Optional[Tuple[int, int]] = None, + callback_on_step_end: Optional[ + Union[Callable[[int, int, Dict], None], PipelineCallback, MultiPipelineCallbacks] + ] = None, + callback_on_step_end_tensor_inputs: List[str] = ["latents"], + max_sequence_length: int = 256, + map: PipelineImageInput = None, + ): + r""" + Function invoked when calling the pipeline for generation. + + Args: + prompt (`str` or `List[str]`, *optional*): + The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`. + instead. + image (`torch.Tensor` or `PIL.Image.Image` or `np.ndarray` or `List[torch.Tensor]` or `List[PIL.Image.Image]` or `List[np.ndarray]`): + The image(s) to modify with the pipeline. + strength (`float`, *optional*, defaults to 0.3): + Conceptually, indicates how much to transform the reference `image`. Must be between 0 and 1. `image` + will be used as a starting point, adding more noise to it the larger the `strength`. The number of + denoising steps depends on the amount of noise initially added. When `strength` is 1, added noise will + be maximum and the denoising process will run for the full number of iterations specified in + `num_inference_steps`. A value of 1, therefore, essentially ignores `image`. Note that in the case of + `denoising_start` being declared as an integer, the value of `strength` will be ignored. + height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor): + The height in pixels of the generated image. This is set to 1024 by default for the best results. + Anything below 512 pixels won't work well for + [Kwai-Kolors/Kolors-diffusers](https://huggingface.co/Kwai-Kolors/Kolors-diffusers) and checkpoints + that are not specifically fine-tuned on low resolutions. + width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor): + The width in pixels of the generated image. This is set to 1024 by default for the best results. + Anything below 512 pixels won't work well for + [Kwai-Kolors/Kolors-diffusers](https://huggingface.co/Kwai-Kolors/Kolors-diffusers) and checkpoints + that are not specifically fine-tuned on low resolutions. + num_inference_steps (`int`, *optional*, defaults to 50): + The number of denoising steps. More denoising steps usually lead to a higher quality image at the + expense of slower inference. + timesteps (`List[int]`, *optional*): + Custom timesteps to use for the denoising process with schedulers which support a `timesteps` argument + in their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is + passed will be used. Must be in descending order. + sigmas (`List[float]`, *optional*): + Custom sigmas to use for the denoising process with schedulers which support a `sigmas` argument in + their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed + will be used. + denoising_start (`float`, *optional*): + When specified, indicates the fraction (between 0.0 and 1.0) of the total denoising process to be + bypassed before it is initiated. Consequently, the initial part of the denoising process is skipped and + it is assumed that the passed `image` is a partly denoised image. Note that when this is specified, + strength will be ignored. The `denoising_start` parameter is particularly beneficial when this pipeline + is integrated into a "Mixture of Denoisers" multi-pipeline setup, as detailed in [**Refine Image + Quality**](https://huggingface.co/docs/diffusers/using-diffusers/sdxl#refine-image-quality). + denoising_end (`float`, *optional*): + When specified, determines the fraction (between 0.0 and 1.0) of the total denoising process to be + completed before it is intentionally prematurely terminated. As a result, the returned sample will + still retain a substantial amount of noise as determined by the discrete timesteps selected by the + scheduler. The denoising_end parameter should ideally be utilized when this pipeline forms a part of a + "Mixture of Denoisers" multi-pipeline setup, as elaborated in [**Refining the Image + Output**](https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/stable_diffusion_xl#refining-the-image-output) + guidance_scale (`float`, *optional*, defaults to 5.0): + Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598). + `guidance_scale` is defined as `w` of equation 2. of [Imagen + Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale > + 1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`, + usually at the expense of lower image quality. + negative_prompt (`str` or `List[str]`, *optional*): + The prompt or prompts not to guide the image generation. If not defined, one has to pass + `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is + less than `1`). + num_images_per_prompt (`int`, *optional*, defaults to 1): + The number of images to generate per prompt. + eta (`float`, *optional*, defaults to 0.0): + Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to + [`schedulers.DDIMScheduler`], will be ignored for others. + generator (`torch.Generator` or `List[torch.Generator]`, *optional*): + One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html) + to make generation deterministic. + latents (`torch.Tensor`, *optional*): + Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image + generation. Can be used to tweak the same generation with different prompts. If not provided, a latents + tensor will ge generated by sampling using the supplied random `generator`. + prompt_embeds (`torch.Tensor`, *optional*): + Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not + provided, text embeddings will be generated from `prompt` input argument. + pooled_prompt_embeds (`torch.Tensor`, *optional*): + Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. + If not provided, pooled text embeddings will be generated from `prompt` input argument. + negative_prompt_embeds (`torch.Tensor`, *optional*): + Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt + weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input + argument. + negative_pooled_prompt_embeds (`torch.Tensor`, *optional*): + Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt + weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt` + input argument. + ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters. + ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*): + Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of + IP-adapters. Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. It should + contain the negative image embedding if `do_classifier_free_guidance` is set to `True`. If not + provided, embeddings are computed from the `ip_adapter_image` input argument. + output_type (`str`, *optional*, defaults to `"pil"`): + The output format of the generate image. Choose between + [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`. + return_dict (`bool`, *optional*, defaults to `True`): + Whether or not to return a [`~pipelines.kolors.KolorsPipelineOutput`] instead of a plain tuple. + cross_attention_kwargs (`dict`, *optional*): + A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under + `self.processor` in + [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py). + original_size (`Tuple[int]`, *optional*, defaults to (1024, 1024)): + If `original_size` is not the same as `target_size` the image will appear to be down- or upsampled. + `original_size` defaults to `(height, width)` if not specified. Part of SDXL's micro-conditioning as + explained in section 2.2 of + [https://huggingface.co/papers/2307.01952](https://huggingface.co/papers/2307.01952). + crops_coords_top_left (`Tuple[int]`, *optional*, defaults to (0, 0)): + `crops_coords_top_left` can be used to generate an image that appears to be "cropped" from the position + `crops_coords_top_left` downwards. Favorable, well-centered images are usually achieved by setting + `crops_coords_top_left` to (0, 0). Part of SDXL's micro-conditioning as explained in section 2.2 of + [https://huggingface.co/papers/2307.01952](https://huggingface.co/papers/2307.01952). + target_size (`Tuple[int]`, *optional*, defaults to (1024, 1024)): + For most cases, `target_size` should be set to the desired height and width of the generated image. If + not specified it will default to `(height, width)`. Part of SDXL's micro-conditioning as explained in + section 2.2 of [https://huggingface.co/papers/2307.01952](https://huggingface.co/papers/2307.01952). + negative_original_size (`Tuple[int]`, *optional*, defaults to (1024, 1024)): + To negatively condition the generation process based on a specific image resolution. Part of SDXL's + micro-conditioning as explained in section 2.2 of + [https://huggingface.co/papers/2307.01952](https://huggingface.co/papers/2307.01952). For more + information, refer to this issue thread: https://github.com/huggingface/diffusers/issues/4208. + negative_crops_coords_top_left (`Tuple[int]`, *optional*, defaults to (0, 0)): + To negatively condition the generation process based on a specific crop coordinates. Part of SDXL's + micro-conditioning as explained in section 2.2 of + [https://huggingface.co/papers/2307.01952](https://huggingface.co/papers/2307.01952). For more + information, refer to this issue thread: https://github.com/huggingface/diffusers/issues/4208. + negative_target_size (`Tuple[int]`, *optional*, defaults to (1024, 1024)): + To negatively condition the generation process based on a target image resolution. It should be as same + as the `target_size` for most cases. Part of SDXL's micro-conditioning as explained in section 2.2 of + [https://huggingface.co/papers/2307.01952](https://huggingface.co/papers/2307.01952). For more + information, refer to this issue thread: https://github.com/huggingface/diffusers/issues/4208. + callback_on_step_end (`Callable`, `PipelineCallback`, `MultiPipelineCallbacks`, *optional*): + A function or a subclass of `PipelineCallback` or `MultiPipelineCallbacks` that is called at the end of + each denoising step during the inference. with the following arguments: `callback_on_step_end(self: + DiffusionPipeline, step: int, timestep: int, callback_kwargs: Dict)`. `callback_kwargs` will include a + list of all tensors as specified by `callback_on_step_end_tensor_inputs`. + callback_on_step_end_tensor_inputs (`List`, *optional*): + The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list + will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the + `._callback_tensor_inputs` attribute of your pipeline class. + max_sequence_length (`int` defaults to 256): Maximum sequence length to use with the `prompt`. + + Examples: + + Returns: + [`~pipelines.kolors.KolorsPipelineOutput`] or `tuple`: [`~pipelines.kolors.KolorsPipelineOutput`] if + `return_dict` is True, otherwise a `tuple`. When returning a tuple, the first element is a list with the + generated images. + """ + + if isinstance(callback_on_step_end, (PipelineCallback, MultiPipelineCallbacks)): + callback_on_step_end_tensor_inputs = callback_on_step_end.tensor_inputs + + # 0. Default height and width to unet + height = height or self.default_sample_size * self.vae_scale_factor + width = width or self.default_sample_size * self.vae_scale_factor + + original_size = original_size or (height, width) + target_size = target_size or (height, width) + + # 1. Check inputs. Raise error if not correct + self.check_inputs( + prompt, + strength, + num_inference_steps, + height, + width, + negative_prompt, + prompt_embeds, + pooled_prompt_embeds, + negative_prompt_embeds, + negative_pooled_prompt_embeds, + ip_adapter_image, + ip_adapter_image_embeds, + callback_on_step_end_tensor_inputs, + max_sequence_length=max_sequence_length, + ) + + self._guidance_scale = guidance_scale + self._cross_attention_kwargs = cross_attention_kwargs + self._denoising_end = denoising_end + self._denoising_start = denoising_start + self._interrupt = False + + # 2. Define call parameters + if prompt is not None and isinstance(prompt, str): + batch_size = 1 + elif prompt is not None and isinstance(prompt, list): + batch_size = len(prompt) + else: + batch_size = prompt_embeds.shape[0] + + device = self._execution_device + + # 3. Encode input prompt + ( + prompt_embeds, + negative_prompt_embeds, + pooled_prompt_embeds, + negative_pooled_prompt_embeds, + ) = self.encode_prompt( + prompt=prompt, + device=device, + num_images_per_prompt=num_images_per_prompt, + do_classifier_free_guidance=self.do_classifier_free_guidance, + negative_prompt=negative_prompt, + prompt_embeds=prompt_embeds, + negative_prompt_embeds=negative_prompt_embeds, + ) + + # 4. Preprocess image + init_image = self.image_processor.preprocess(image, height=height, width=width).to(dtype=torch.float32) + + map = self.mask_processor.preprocess( + map, height=height // self.vae_scale_factor, width=width // self.vae_scale_factor + ).to(device) + + # 5. Prepare timesteps + def denoising_value_valid(dnv): + return isinstance(dnv, float) and 0 < dnv < 1 + + timesteps, num_inference_steps = retrieve_timesteps( + self.scheduler, num_inference_steps, device, timesteps, sigmas + ) + + # begin diff diff change + total_time_steps = num_inference_steps + # end diff diff change + + timesteps, num_inference_steps = self.get_timesteps( + num_inference_steps, + strength, + device, + denoising_start=self.denoising_start if denoising_value_valid(self.denoising_start) else None, + ) + latent_timestep = timesteps[:1].repeat(batch_size * num_images_per_prompt) + + add_noise = True if self.denoising_start is None else False + + # 6. Prepare latent variables + if latents is None: + latents = self.prepare_latents( + init_image, + latent_timestep, + batch_size, + num_images_per_prompt, + prompt_embeds.dtype, + device, + generator, + add_noise, + ) + + # 7. Prepare extra step kwargs. + extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta) + + height, width = latents.shape[-2:] + height = height * self.vae_scale_factor + width = width * self.vae_scale_factor + + original_size = original_size or (height, width) + target_size = target_size or (height, width) + + # 8. Prepare added time ids & embeddings + add_text_embeds = pooled_prompt_embeds + text_encoder_projection_dim = int(pooled_prompt_embeds.shape[-1]) + + add_time_ids = self._get_add_time_ids( + original_size, + crops_coords_top_left, + target_size, + dtype=prompt_embeds.dtype, + text_encoder_projection_dim=text_encoder_projection_dim, + ) + if negative_original_size is not None and negative_target_size is not None: + negative_add_time_ids = self._get_add_time_ids( + negative_original_size, + negative_crops_coords_top_left, + negative_target_size, + dtype=prompt_embeds.dtype, + text_encoder_projection_dim=text_encoder_projection_dim, + ) + else: + negative_add_time_ids = add_time_ids + + if self.do_classifier_free_guidance: + prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds], dim=0) + add_text_embeds = torch.cat([negative_pooled_prompt_embeds, add_text_embeds], dim=0) + add_time_ids = torch.cat([negative_add_time_ids, add_time_ids], dim=0) + + prompt_embeds = prompt_embeds.to(device) + add_text_embeds = add_text_embeds.to(device) + add_time_ids = add_time_ids.to(device).repeat(batch_size * num_images_per_prompt, 1) + + if ip_adapter_image is not None or ip_adapter_image_embeds is not None: + image_embeds = self.prepare_ip_adapter_image_embeds( + ip_adapter_image, + ip_adapter_image_embeds, + device, + batch_size * num_images_per_prompt, + self.do_classifier_free_guidance, + ) + + # 9. Denoising loop + num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0) + + # preparations for diff diff + original_with_noise = self.prepare_latents( + init_image, timesteps, batch_size, num_images_per_prompt, prompt_embeds.dtype, device, generator + ) + thresholds = torch.arange(total_time_steps, dtype=map.dtype) / total_time_steps + thresholds = thresholds.unsqueeze(1).unsqueeze(1).to(device) + masks = map.squeeze() > thresholds + # end diff diff preparations + + # 9.1 Apply denoising_end + if ( + self.denoising_end is not None + and self.denoising_start is not None + and denoising_value_valid(self.denoising_end) + and denoising_value_valid(self.denoising_start) + and self.denoising_start >= self.denoising_end + ): + raise ValueError( + f"`denoising_start`: {self.denoising_start} cannot be larger than or equal to `denoising_end`: " + + f" {self.denoising_end} when using type float." + ) + elif self.denoising_end is not None and denoising_value_valid(self.denoising_end): + discrete_timestep_cutoff = int( + round( + self.scheduler.config.num_train_timesteps + - (self.denoising_end * self.scheduler.config.num_train_timesteps) + ) + ) + num_inference_steps = len(list(filter(lambda ts: ts >= discrete_timestep_cutoff, timesteps))) + timesteps = timesteps[:num_inference_steps] + + # 9.2 Optionally get Guidance Scale Embedding + timestep_cond = None + if self.unet.config.time_cond_proj_dim is not None: + guidance_scale_tensor = torch.tensor(self.guidance_scale - 1).repeat(batch_size * num_images_per_prompt) + timestep_cond = self.get_guidance_scale_embedding( + guidance_scale_tensor, embedding_dim=self.unet.config.time_cond_proj_dim + ).to(device=device, dtype=latents.dtype) + + self._num_timesteps = len(timesteps) + with self.progress_bar(total=num_inference_steps) as progress_bar: + for i, t in enumerate(timesteps): + if self.interrupt: + continue + + # diff diff + if i == 0: + latents = original_with_noise[:1] + else: + mask = masks[i].unsqueeze(0).to(latents.dtype) + mask = mask.unsqueeze(1) # fit shape + latents = original_with_noise[i] * mask + latents * (1 - mask) + # end diff diff + + # expand the latents if we are doing classifier free guidance + latent_model_input = torch.cat([latents] * 2) if self.do_classifier_free_guidance else latents + + latent_model_input = self.scheduler.scale_model_input(latent_model_input, t) + + # predict the noise residual + added_cond_kwargs = {"text_embeds": add_text_embeds, "time_ids": add_time_ids} + + if ip_adapter_image is not None or ip_adapter_image_embeds is not None: + added_cond_kwargs["image_embeds"] = image_embeds + + noise_pred = self.unet( + latent_model_input, + t, + encoder_hidden_states=prompt_embeds, + timestep_cond=timestep_cond, + cross_attention_kwargs=self.cross_attention_kwargs, + added_cond_kwargs=added_cond_kwargs, + return_dict=False, + )[0] + + # perform guidance + if self.do_classifier_free_guidance: + noise_pred_uncond, noise_pred_text = noise_pred.chunk(2) + noise_pred = noise_pred_uncond + self.guidance_scale * (noise_pred_text - noise_pred_uncond) + + # compute the previous noisy sample x_t -> x_t-1 + latents_dtype = latents.dtype + latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs, return_dict=False)[0] + if latents.dtype != latents_dtype: + if torch.backends.mps.is_available(): + # some platforms (eg. apple mps) misbehave due to a pytorch bug: https://github.com/pytorch/pytorch/pull/99272 + latents = latents.to(latents_dtype) + + if callback_on_step_end is not None: + callback_kwargs = {} + for k in callback_on_step_end_tensor_inputs: + callback_kwargs[k] = locals()[k] + callback_outputs = callback_on_step_end(self, i, t, callback_kwargs) + + latents = callback_outputs.pop("latents", latents) + prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds) + negative_prompt_embeds = callback_outputs.pop("negative_prompt_embeds", negative_prompt_embeds) + add_text_embeds = callback_outputs.pop("add_text_embeds", add_text_embeds) + negative_pooled_prompt_embeds = callback_outputs.pop( + "negative_pooled_prompt_embeds", negative_pooled_prompt_embeds + ) + add_time_ids = callback_outputs.pop("add_time_ids", add_time_ids) + negative_add_time_ids = callback_outputs.pop("negative_add_time_ids", negative_add_time_ids) + + if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0): + progress_bar.update() + + if XLA_AVAILABLE: + xm.mark_step() + + if not output_type == "latent": + # make sure the VAE is in float32 mode, as it overflows in float16 + needs_upcasting = self.vae.dtype == torch.float16 and self.vae.config.force_upcast + + if needs_upcasting: + self.upcast_vae() + latents = latents.to(next(iter(self.vae.post_quant_conv.parameters())).dtype) + elif latents.dtype != self.vae.dtype: + if torch.backends.mps.is_available(): + # some platforms (eg. apple mps) misbehave due to a pytorch bug: https://github.com/pytorch/pytorch/pull/99272 + self.vae = self.vae.to(latents.dtype) + + # unscale/denormalize the latents + latents = latents / self.vae.config.scaling_factor + + image = self.vae.decode(latents, return_dict=False)[0] + + # cast back to fp16 if needed + if needs_upcasting: + self.vae.to(dtype=torch.float16) + else: + image = latents + + if not output_type == "latent": + image = self.image_processor.postprocess(image, output_type=output_type) + + # Offload all models + self.maybe_free_model_hooks() + + if not return_dict: + return (image,) + + return KolorsPipelineOutput(images=image) From 0ce456b5d7f5adc5dac2a92a90a5b773ac4e227e Mon Sep 17 00:00:00 2001 From: hlky Date: Fri, 11 Oct 2024 18:39:19 +0100 Subject: [PATCH 09/12] FluxMultiControlNetModel (#9647) --- src/diffusers/pipelines/flux/pipeline_flux_controlnet.py | 2 ++ .../pipelines/flux/pipeline_flux_controlnet_image_to_image.py | 2 ++ .../pipelines/flux/pipeline_flux_controlnet_inpainting.py | 2 ++ 3 files changed, 6 insertions(+) diff --git a/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py b/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py index a301f6742c05..8770c231809f 100644 --- a/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py +++ b/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py @@ -202,6 +202,8 @@ def __init__( ], ): super().__init__() + if isinstance(controlnet, (list, tuple)): + controlnet = FluxMultiControlNetModel(controlnet) self.register_modules( vae=vae, diff --git a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py index b5ff2236c4d0..a7f7c66a2cad 100644 --- a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py +++ b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py @@ -214,6 +214,8 @@ def __init__( ], ): super().__init__() + if isinstance(controlnet, (list, tuple)): + controlnet = FluxMultiControlNetModel(controlnet) self.register_modules( vae=vae, diff --git a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py index 1c1c25302100..50d2fcaa7fa5 100644 --- a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py +++ b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py @@ -216,6 +216,8 @@ def __init__( ], ): super().__init__() + if isinstance(controlnet, (list, tuple)): + controlnet = FluxMultiControlNetModel(controlnet) self.register_modules( scheduler=scheduler, From 151ec123bf817115133144527d9a2742fd23b883 Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Sat, 12 Oct 2024 02:10:38 +0530 Subject: [PATCH 10/12] tests --- tests/lora/test_lora_layers_flux.py | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-) diff --git a/tests/lora/test_lora_layers_flux.py b/tests/lora/test_lora_layers_flux.py index 4629c24c8cd8..3bc46d1e9b13 100644 --- a/tests/lora/test_lora_layers_flux.py +++ b/tests/lora/test_lora_layers_flux.py @@ -27,6 +27,7 @@ from diffusers.utils.testing_utils import ( floats_tensor, is_peft_available, + numpy_cosine_similarity_distance, require_peft_backend, require_torch_gpu, slow, @@ -166,7 +167,7 @@ def test_modify_padding_mode(self): @slow @require_torch_gpu @require_peft_backend -@unittest.skip("We cannot run inference on this model with the current CI hardware") +# @unittest.skip("We cannot run inference on this model with the current CI hardware") # TODO (DN6, sayakpaul): move these tests to a beefier GPU class FluxLoRAIntegrationTests(unittest.TestCase): """internal note: The integration slices were obtained on audace. @@ -208,9 +209,11 @@ def test_flux_the_last_ben(self): generator=torch.manual_seed(self.seed), ).images out_slice = out[0, -3:, -3:, -1].flatten() - expected_slice = np.array([0.1719, 0.1719, 0.1699, 0.1719, 0.1719, 0.1738, 0.1641, 0.1621, 0.2090]) + expected_slice = np.array([0.1855, 0.1855, 0.1836, 0.1855, 0.1836, 0.1875, 0.1777, 0.1758, 0.2246]) - assert np.allclose(out_slice, expected_slice, atol=1e-4, rtol=1e-4) + max_diff = numpy_cosine_similarity_distance(expected_slice.flatten(), out_slice) + + assert max_diff < 1e-3 def test_flux_kohya(self): self.pipeline.load_lora_weights("Norod78/brain-slug-flux") @@ -230,7 +233,9 @@ def test_flux_kohya(self): out_slice = out[0, -3:, -3:, -1].flatten() expected_slice = np.array([0.6367, 0.6367, 0.6328, 0.6367, 0.6328, 0.6289, 0.6367, 0.6328, 0.6484]) - assert np.allclose(out_slice, expected_slice, atol=1e-4, rtol=1e-4) + max_diff = numpy_cosine_similarity_distance(expected_slice.flatten(), out_slice) + + assert max_diff < 1e-3 def test_flux_kohya_with_text_encoder(self): self.pipeline.load_lora_weights("cocktailpeanut/optimus", weight_name="optimus.safetensors") @@ -248,9 +253,11 @@ def test_flux_kohya_with_text_encoder(self): ).images out_slice = out[0, -3:, -3:, -1].flatten() - expected_slice = np.array([0.4023, 0.4043, 0.4023, 0.3965, 0.3984, 0.3984, 0.3906, 0.3906, 0.4219]) + expected_slice = np.array([0.4023, 0.4023, 0.4023, 0.3965, 0.3984, 0.3965, 0.3926, 0.3906, 0.4219]) - assert np.allclose(out_slice, expected_slice, atol=1e-4, rtol=1e-4) + max_diff = numpy_cosine_similarity_distance(expected_slice.flatten(), out_slice) + + assert max_diff < 1e-3 def test_flux_xlabs(self): self.pipeline.load_lora_weights("XLabs-AI/flux-lora-collection", weight_name="disney_lora.safetensors") @@ -268,6 +275,8 @@ def test_flux_xlabs(self): generator=torch.manual_seed(self.seed), ).images out_slice = out[0, -3:, -3:, -1].flatten() - expected_slice = np.array([0.3984, 0.4199, 0.4453, 0.4102, 0.4375, 0.4590, 0.4141, 0.4355, 0.4980]) + expected_slice = np.array([0.3965, 0.4180, 0.4434, 0.4082, 0.4375, 0.4590, 0.4141, 0.4375, 0.4980]) + + max_diff = numpy_cosine_similarity_distance(expected_slice.flatten(), out_slice) - assert np.allclose(out_slice, expected_slice, atol=1e-4, rtol=1e-4) + assert max_diff < 1e-3 From dfd9865c5ce465790b490895c05dde833c3de902 Mon Sep 17 00:00:00 2001 From: Sayak Paul Date: Tue, 15 Oct 2024 08:55:09 +0530 Subject: [PATCH 11/12] Update src/diffusers/loaders/lora_pipeline.py Co-authored-by: YiYi Xu --- src/diffusers/loaders/lora_pipeline.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index 39414c122c26..dc42e5c67770 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -1366,7 +1366,7 @@ def load_lora_into_transformer( lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: warn_msg = ( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f"Loading adapter weights from state_dict led to unexpected keys found in the model:" f" {', '.join(lora_unexpected_keys)}. " ) From f197a37c12ded65dd58ebc7a44a94f81b4d744ef Mon Sep 17 00:00:00 2001 From: sayakpaul Date: Tue, 15 Oct 2024 08:58:26 +0530 Subject: [PATCH 12/12] fix --- src/diffusers/loaders/lora_pipeline.py | 6 +++--- src/diffusers/loaders/unet.py | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/src/diffusers/loaders/lora_pipeline.py b/src/diffusers/loaders/lora_pipeline.py index dc42e5c67770..5e01ec567f9a 100644 --- a/src/diffusers/loaders/lora_pipeline.py +++ b/src/diffusers/loaders/lora_pipeline.py @@ -1956,7 +1956,7 @@ def load_lora_into_transformer( lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: warn_msg = ( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f"Loading adapter weights from state_dict led to unexpected keys found in the model:" f" {', '.join(lora_unexpected_keys)}. " ) @@ -2319,7 +2319,7 @@ def load_lora_into_transformer(cls, state_dict, network_alphas, transformer, ada lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: warn_msg = ( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f"Loading adapter weights from state_dict led to unexpected keys found in the model:" f" {', '.join(lora_unexpected_keys)}. " ) @@ -2773,7 +2773,7 @@ def load_lora_into_transformer( lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: warn_msg = ( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f"Loading adapter weights from state_dict led to unexpected keys found in the model:" f" {', '.join(lora_unexpected_keys)}. " ) diff --git a/src/diffusers/loaders/unet.py b/src/diffusers/loaders/unet.py index 8f302bdad31b..2fa7732a6a3b 100644 --- a/src/diffusers/loaders/unet.py +++ b/src/diffusers/loaders/unet.py @@ -362,7 +362,7 @@ def _process_lora( lora_unexpected_keys = [k for k in unexpected_keys if "lora_" in k and adapter_name in k] if lora_unexpected_keys: warn_msg = ( - f"Loading adapter weights from state_dict led to unexpected keys not found in the model:" + f"Loading adapter weights from state_dict led to unexpected keys found in the model:" f" {', '.join(lora_unexpected_keys)}. " )