runs on mac but generation is wrong

app.py

@@ -0,0 +1,263 @@
+import gradio as gr
+import torch
+from PIL import Image
+from torchvision import transforms
+from diffusers import StableDiffusionPipeline, StableDiffusionImageVariationPipeline, DiffusionPipeline
+import numpy as np
+import pandas as pd
+import math
+from transformers import CLIPTextModel, CLIPTokenizer
+
+# model_id = "stabilityai/stable-diffusion-2-1-base"
+# text_model_id = "CompVis/stable-diffusion-v-1-4-original"
+# text_model_id = "CompVis/stable-diffusion-v1-4"
+text_model_id = "runwayml/stable-diffusion-v1-5"
+# text_model_id = "stabilityai/stable-diffusion-2-1-base"
+model_id = "lambdalabs/sd-image-variations-diffusers"
+clip_model_id = "openai/clip-vit-large-patch14-336"
+
+max_tabs = 5
+input_images = [None for i in range(max_tabs)]
+input_prompts = [None for i in range(max_tabs)]
+embedding_plots = [None for i in range(max_tabs)]
+# global embedding_base64s
+embedding_base64s = [None for i in range(max_tabs)]
+# embedding_base64s = gr.State(value=[None for i in range(max_tabs)])
+
+
+def image_to_embedding(input_im):
+    tform = transforms.Compose([
+    transforms.ToTensor(),
+    transforms.Resize(
+        (224, 224),
+        interpolation=transforms.InterpolationMode.BICUBIC,
+        antialias=False,
+        ),
+        transforms.Normalize(
+          [0.48145466, 0.4578275, 0.40821073],
+          [0.26862954, 0.26130258, 0.27577711]),
+    ])
+
+    inp = tform(input_im).to(device)
+    dtype = next(pipe.image_encoder.parameters()).dtype
+    image = inp.tile(1, 1, 1, 1).to(device=device, dtype=dtype)
+    image_embeddings = pipe.image_encoder(image).image_embeds
+    image_embeddings = image_embeddings[0]
+    image_embeddings_np = image_embeddings.cpu().detach().numpy()
+    return image_embeddings_np
+
+def prompt_to_embedding(prompt):
+    # inputs = processor(prompt, images=imgs, return_tensors="pt", padding=True)
+    inputs = processor(prompt, return_tensors="pt", padding='max_length', max_length=77)
+    # labels = torch.tensor(labels)
+    # prompt_tokens = inputs.input_ids[0]
+    prompt_tokens = inputs.input_ids
+    # image = inputs.pixel_values
+    with torch.no_grad():
+        prompt_embededdings = model.get_text_features(prompt_tokens.to(device))
+    prompt_embededdings = prompt_embededdings[0].cpu().detach().numpy()
+    return prompt_embededdings
+
+def embedding_to_image(embeddings):
+    size = math.ceil(math.sqrt(embeddings.shape[0]))
+    image_embeddings_square = np.pad(embeddings, (0, size**2 - embeddings.shape[0]), 'constant')
+    image_embeddings_square.resize(size,size)
+    embedding_image = Image.fromarray(image_embeddings_square, mode="L")
+    return embedding_image
+
+def embedding_to_base64(embeddings):
+    import base64
+    # ensure float16
+    embeddings = embeddings.astype(np.float16)
+    embeddings_b64 = base64.urlsafe_b64encode(embeddings).decode()
+    return embeddings_b64
+
+def base64_to_embedding(embeddings_b64):
+    import base64
+    embeddings = base64.urlsafe_b64decode(embeddings_b64)
+    embeddings = np.frombuffer(embeddings, dtype=np.float16)
+    # embeddings = torch.tensor(embeddings)
+    return embeddings
+
+def main(
+    # input_im,
+    embeddings,
+    scale=3.0,
+    n_samples=4,
+    steps=25,
+    seed=0
+    ):
+
+    if seed == None:
+        seed = np.random.randint(2147483647)
+    # generator = torch.Generator(device=device).manual_seed(int(seed))
+    generator = torch.Generator().manual_seed(int(seed)) # use cpu as does not work on mps
+
+    embeddings = base64_to_embedding(embeddings)
+    embeddings = torch.tensor(embeddings).to(device)
+
+    images_list = pipe(
+        # inp.tile(n_samples, 1, 1, 1),
+        # [embeddings * n_samples],
+        embeddings,
+        guidance_scale=scale,
+        num_inference_steps=steps,
+        generator=generator,
+        )
+
+    images = []
+    for i, image in enumerate(images_list["images"]):
+        images.append(image)
+    # images.append(embedding_image)
+    return images
+
+def on_image_load_update_embeddings(image_data):
+    # image to embeddings
+    if image_data is None:
+        embeddings = prompt_to_embedding('')
+        embeddings_b64 = embedding_to_base64(embeddings)
+        return gr.Text.update(embeddings_b64)
+    embeddings = image_to_embedding(image_data)
+    embeddings_b64 = embedding_to_base64(embeddings)
+    return gr.Text.update(embeddings_b64)
+
+def on_prompt_change_update_embeddings(prompt):
+    # prompt to embeddings
+    if prompt is None or prompt == "":
+        embeddings = prompt_to_embedding('')
+        embeddings_b64 = embedding_to_base64(embeddings)
+        return gr.Text.update(embedding_to_base64(embeddings))
+    embeddings = prompt_to_embedding(prompt)
+    embeddings_b64 = embedding_to_base64(embeddings)
+    return gr.Text.update(embeddings_b64)
+
+# def on_embeddings_changed_update_average_embeddings(last_embedding_base64):
+# def on_embeddings_changed_update_average_embeddings(embedding_base64s):
+def on_embeddings_changed_update_average_embeddings(embedding_base64s_state, embedding_base64, idx):
+    # global embedding_base64s
+    final_embedding = None
+    num_embeddings = 0
+    embedding_base64s_state[idx] = embedding_base64
+    # for textbox in embedding_base64s:
+    #     embedding_base64 = textbox.value
+    for embedding_base64 in embedding_base64s_state:
+        if embedding_base64 is None or embedding_base64 == "":
+            continue
+        embedding = base64_to_embedding(embedding_base64)
+        if final_embedding is None:
+            final_embedding = embedding
+        else:
+            final_embedding = final_embedding + embedding
+        num_embeddings += 1
+    if final_embedding is None:
+        embeddings = prompt_to_embedding('')
+        embeddings_b64 = embedding_to_base64(embeddings)
+        return gr.Text.update(embeddings_b64)
+    final_embedding = final_embedding / num_embeddings
+    embeddings_b64 = embedding_to_base64(final_embedding)
+    return gr.Text.update(embeddings_b64)    
+
+def on_embeddings_changed_update_plot(embeddings_b64):
+    # plot new embeddings
+    if embeddings_b64 is None or embeddings_b64 == "":
+        return gr.LinePlot.update()
+        
+    embeddings = base64_to_embedding(embeddings_b64)
+    data = pd.DataFrame({
+            'embedding': embeddings,
+            'index': [n for n in range(len(embeddings))]})
+    return gr.LinePlot.update(data,
+            x="index",
+            y="embedding",
+            # color="country",
+            title="Embeddings",
+            # stroke_dash="cluster",
+            # x_lim=[1950, 2010],
+            tooltip=['index', 'embedding'],
+            # stroke_dash_legend_title="Country Cluster",
+            # height=300,
+            width=embeddings.shape[0])
+
+
+device = torch.device("mps" if torch.backends.mps.is_available() else "cuda:0" if torch.cuda.is_available() else "cpu")
+pipe = StableDiffusionPipeline.from_pretrained(
+    model_id,
+    custom_pipeline="pipeline.py",
+    torch_dtype=torch.float16,
+    # , revision="fp16",
+    requires_safety_checker = False, safety_checker=None,
+    text_encoder = CLIPTextModel,
+    tokenizer = CLIPTokenizer,
+    )
+pipe = pipe.to(device)
+
+from transformers import AutoProcessor, AutoModel
+processor = AutoProcessor.from_pretrained(clip_model_id)
+model = AutoModel.from_pretrained(clip_model_id)
+model = model.to(device)
+
+examples = [
+    ["frog.png", 3, 1, 25, 0],
+    ["img0.jpg", 3, 1, 25, 0],
+    ["img1.jpg", 3, 1, 25, 0],
+    ["img2.jpg", 3, 1, 25, 0],
+    ["img3.jpg", 3, 1, 25, 0],
+]
+
+
+with gr.Blocks() as demo:
+    with gr.Row():
+        for i in range(max_tabs):
+            with gr.Tab(f"Input {i}"):
+                with gr.Row():
+                    with gr.Column(scale=1, min_width=240):
+                        input_images[i] = gr.Image()
+                    with gr.Column(scale=3, min_width=600):
+                        embedding_plots[i] = gr.LinePlot(show_label=False).style(container=False)
+                        # input_image.change(on_image_load, inputs= [input_image, plot])
+                with gr.Row():
+                    with gr.Column(scale=1, min_width=240):
+                        input_prompts[i] = gr.Textbox()
+                    with gr.Column(scale=3, min_width=600):
+                        with gr.Accordion("Embeddings", open=False):
+                            embedding_base64s[i] = gr.Textbox(show_label=False)
+
+    with gr.Row():
+        average_embedding_plot = gr.LinePlot(show_label=False).style(container=False)
+    with gr.Row():
+        average_embedding_base64 = gr.Textbox(show_label=False)
+
+    with gr.Row():
+        with gr.Column(scale=1, min_width=200):
+            scale = gr.Slider(0, 25, value=3, step=1, label="Guidance scale")
+        with gr.Column(scale=1, min_width=200):
+            n_samples = gr.Slider(1, 4, value=1, step=1, label="Number images")
+        with gr.Column(scale=1, min_width=200):
+            steps = gr.Slider(5, 50, value=25, step=5, label="Steps")
+        with gr.Column(scale=1, min_width=200):
+            seed = gr.Number(None, label="Seed", precision=0)
+    with gr.Row():
+        submit = gr.Button("Submit")
+    with gr.Row():
+        output = gr.Gallery(label="Generated variations")
+
+    embedding_base64s_state = gr.State(value=[None for i in range(max_tabs)])
+    for i in range(max_tabs):
+        input_images[i].change(on_image_load_update_embeddings, input_images[i], [embedding_base64s[i]])
+        input_prompts[i].submit(on_prompt_change_update_embeddings, input_prompts[i], [embedding_base64s[i]])
+        embedding_base64s[i].change(on_embeddings_changed_update_plot, embedding_base64s[i], [embedding_plots[i]])
+        # embedding_plots[i].change(on_plot_changed, embedding_base64s[i], average_embedding_base64)
+        # embedding_plots[i].change(on_embeddings_changed_update_average_embeddings, embedding_base64s[i], average_embedding_base64)
+        idx_state = gr.State(value=i)
+        embedding_base64s[i].change(on_embeddings_changed_update_average_embeddings, [embedding_base64s_state, embedding_base64s[i], idx_state], average_embedding_base64)
+
+    average_embedding_base64.change(on_embeddings_changed_update_plot, average_embedding_base64, average_embedding_plot)
+
+    # submit.click(main, inputs= [embedding_base64s[0], scale, n_samples, steps, seed], outputs=output)
+    submit.click(main, inputs= [average_embedding_base64, scale, n_samples, steps, seed], outputs=output)
+    output.style(grid=2)
+
+    
+
+if __name__ == "__main__":
+    demo.launch() 

pipeline.py

@@ -0,0 +1,486 @@
+# Copyright 2022 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 Callable, List, Optional, Union
+
+import torch
+
+import PIL
+from diffusers.utils import is_accelerate_available
+from packaging import version
+from transformers import CLIPFeatureExtractor, CLIPVisionModelWithProjection
+from transformers import CLIPTextModel, CLIPTokenizer
+
+# from ...configuration_utils import FrozenDict
+# from ...models import AutoencoderKL, UNet2DConditionModel
+# from ...pipeline_utils import DiffusionPipeline
+# from ...schedulers import (
+#     DDIMScheduler,
+#     DPMSolverMultistepScheduler,
+#     EulerAncestralDiscreteScheduler,
+#     EulerDiscreteScheduler,
+#     LMSDiscreteScheduler,
+#     PNDMScheduler,
+# )
+# from ...utils import deprecate, logging
+# from . import StableDiffusionPipelineOutput
+# from .safety_checker import StableDiffusionSafetyChecker
+from diffusers.configuration_utils import FrozenDict
+from diffusers.models import AutoencoderKL, UNet2DConditionModel
+from diffusers.pipeline_utils import DiffusionPipeline
+from diffusers.schedulers import (
+    DDIMScheduler,
+    DPMSolverMultistepScheduler,
+    EulerAncestralDiscreteScheduler,
+    EulerDiscreteScheduler,
+    LMSDiscreteScheduler,
+    PNDMScheduler,
+)
+from diffusers.utils import deprecate, logging
+from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput, StableDiffusionSafetyChecker
+
+
+logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
+
+
+class StableDiffusionImageTextVariationPipeline(DiffusionPipeline):
+    r"""
+    Pipeline to generate variations from an input image using Stable Diffusion.
+
+    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.)
+
+    Args:
+        vae ([`AutoencoderKL`]):
+            Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
+        image_encoder ([`CLIPVisionModelWithProjection`]):
+            Frozen CLIP image-encoder. Stable Diffusion Image Variation uses the vision portion of
+            [CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPVisionModelWithProjection),
+            specifically the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
+        text_encoder ([`CLIPTextModel`]):
+            Frozen text-encoder. Stable Diffusion uses the text portion of
+            [CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
+            the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
+        tokenizer (`CLIPTokenizer`):
+            Tokenizer of class
+            [CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
+        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`].
+        safety_checker ([`StableDiffusionSafetyChecker`]):
+            Classification module that estimates whether generated images could be considered offensive or harmful.
+            Please, refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for details.
+        feature_extractor ([`CLIPFeatureExtractor`]):
+            Model that extracts features from generated images to be used as inputs for the `safety_checker`.
+    """
+    _optional_components = ["safety_checker"]
+
+    def __init__(
+        self,
+        vae: AutoencoderKL,
+        image_encoder: CLIPVisionModelWithProjection,
+        text_encoder: CLIPTextModel,
+        tokenizer: CLIPTokenizer,
+        unet: UNet2DConditionModel,
+        scheduler: Union[
+            DDIMScheduler,
+            PNDMScheduler,
+            LMSDiscreteScheduler,
+            EulerDiscreteScheduler,
+            EulerAncestralDiscreteScheduler,
+            DPMSolverMultistepScheduler,
+        ],
+        safety_checker: StableDiffusionSafetyChecker,
+        feature_extractor: CLIPFeatureExtractor,
+        requires_safety_checker: bool = True,
+    ):
+        super().__init__()
+
+        if safety_checker is None and requires_safety_checker:
+            logger.warn(
+                f"You have disabled the safety checker for {self.__class__} by passing `safety_checker=None`. Ensure"
+                " that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered"
+                " results in services or applications open to the public. Both the diffusers team and Hugging Face"
+                " strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling"
+                " it only for use-cases that involve analyzing network behavior or auditing its results. For more"
+                " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
+            )
+
+        if safety_checker is not None and feature_extractor is None:
+            raise ValueError(
+                "Make sure to define a feature extractor when loading {self.__class__} if you want to use the safety"
+                " checker. If you do not want to use the safety checker, you can pass `'safety_checker=None'` instead."
+            )
+
+        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
+            version.parse(unet.config._diffusers_version).base_version
+        ) < version.parse("0.9.0.dev0")
+        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
+            deprecation_message = (
+                "The configuration file of the unet has set the default `sample_size` to smaller than"
+                " 64 which seems highly unlikely .If you're checkpoint is a fine-tuned version of any of the"
+                " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
+                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
+                " in the config might lead to incorrect results in future versions. If you have downloaded this"
+                " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
+                " the `unet/config.json` file"
+            )
+            deprecate("sample_size<64", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(unet.config)
+            new_config["sample_size"] = 64
+            unet._internal_dict = FrozenDict(new_config)
+
+        self.register_modules(
+            vae=vae,
+            image_encoder=image_encoder,
+            text_encoder=text_encoder,
+            tokenizer=tokenizer,
+            unet=unet,
+            scheduler=scheduler,
+            safety_checker=safety_checker,
+            feature_extractor=feature_extractor,
+        )
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.register_to_config(requires_safety_checker=requires_safety_checker)
+
+    def enable_sequential_cpu_offload(self, gpu_id=0):
+        r"""
+        Offloads all models to CPU using accelerate, significantly reducing memory usage. When called, unet,
+        text_encoder, vae and safety checker have their state dicts saved to CPU and then are moved to a
+        `torch.device('meta') and loaded to GPU only when their specific submodule has its `forward` method called.
+        """
+        if is_accelerate_available():
+            from accelerate import cpu_offload
+        else:
+            raise ImportError("Please install accelerate via `pip install accelerate`")
+
+        device = torch.device(f"cuda:{gpu_id}")
+
+        for cpu_offloaded_model in [self.unet, self.image_encoder, self.vae, self.safety_checker]:
+            if cpu_offloaded_model is not None:
+                cpu_offload(cpu_offloaded_model, device)
+
+    @property
+    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._execution_device
+    def _execution_device(self):
+        r"""
+        Returns the device on which the pipeline's models will be executed. After calling
+        `pipeline.enable_sequential_cpu_offload()` the execution device can only be inferred from Accelerate's module
+        hooks.
+        """
+        if self.device != torch.device("meta") or not hasattr(self.unet, "_hf_hook"):
+            return self.device
+        for module in self.unet.modules():
+            if (
+                hasattr(module, "_hf_hook")
+                and hasattr(module._hf_hook, "execution_device")
+                and module._hf_hook.execution_device is not None
+            ):
+                return torch.device(module._hf_hook.execution_device)
+        return self.device
+
+    def _encode_image(self, image, device, num_images_per_prompt, do_classifier_free_guidance):
+        dtype = next(self.image_encoder.parameters()).dtype
+
+        if not isinstance(image, torch.Tensor):
+            image = self.feature_extractor(images=image, return_tensors="pt").pixel_values
+
+        image = image.to(device=device, dtype=dtype)
+        image_embeddings = self.image_encoder(image).image_embeds
+        image_embeddings = image_embeddings.unsqueeze(1)
+
+        # duplicate image embeddings for each generation per prompt, using mps friendly method
+        bs_embed, seq_len, _ = image_embeddings.shape
+        image_embeddings = image_embeddings.repeat(1, num_images_per_prompt, 1)
+        image_embeddings = image_embeddings.view(bs_embed * num_images_per_prompt, seq_len, -1)
+
+        if do_classifier_free_guidance:
+            uncond_embeddings = torch.zeros_like(image_embeddings)
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            image_embeddings = torch.cat([uncond_embeddings, image_embeddings])
+
+        return image_embeddings
+
+    def _prepare_embeddings(self, embeddings, device, num_images_per_prompt, do_classifier_free_guidance):
+        dtype = next(self.image_encoder.parameters()).dtype
+
+        # duplicate image embeddings for each generation per prompt, using mps friendly method
+        # bs_embed, seq_len, _ = embeddings.shape
+        # bs_embed = len(embeddings)
+        # seq_len = embeddings[0].shape[0]
+        # embeddings = embeddings.repeat(1, num_images_per_prompt, 1)
+        embeddings = embeddings.repeat(1, 1, num_images_per_prompt)
+        # embeddings = embeddings.view(bs_embed * num_images_per_prompt, seq_len, -1)
+
+        if do_classifier_free_guidance:
+            uncond_embeddings = torch.zeros_like(embeddings)
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            embeddings = torch.cat([uncond_embeddings, embeddings])
+
+        return embeddings        
+
+    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.run_safety_checker
+    def run_safety_checker(self, image, device, dtype):
+        if self.safety_checker is not None:
+            safety_checker_input = self.feature_extractor(self.numpy_to_pil(image), return_tensors="pt").to(device)
+            image, has_nsfw_concept = self.safety_checker(
+                images=image, clip_input=safety_checker_input.pixel_values.to(dtype)
+            )
+        else:
+            has_nsfw_concept = None
+        return image, has_nsfw_concept
+
+    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.decode_latents
+    def decode_latents(self, latents):
+        latents = 1 / 0.18215 * latents
+        image = self.vae.decode(latents).sample
+        image = (image / 2 + 0.5).clamp(0, 1)
+        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloa16
+        image = image.cpu().permute(0, 2, 3, 1).float().numpy()
+        return image
+
+    # 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, image, height, width, callback_steps):
+        if (
+            not isinstance(image, torch.Tensor)
+            and not isinstance(image, PIL.Image.Image)
+            and not isinstance(image, list)
+        ):
+            raise ValueError(
+                "`image` has to be of type `torch.FloatTensor` or `PIL.Image.Image` or `List[PIL.Image.Image]` but is"
+                f" {type(image)}"
+            )
+
+        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_steps is None) or (
+            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
+        ):
+            raise ValueError(
+                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
+                f" {type(callback_steps)}."
+            )
+
+    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents
+    def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype, device, generator, latents=None):
+        shape = (batch_size, num_channels_latents, height // self.vae_scale_factor, width // self.vae_scale_factor)
+        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."
+            )
+
+        if latents is None:
+            rand_device = "cpu" if device.type == "mps" else device
+
+            if isinstance(generator, list):
+                shape = (1,) + shape[1:]
+                latents = [
+                    torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype)
+                    for i in range(batch_size)
+                ]
+                latents = torch.cat(latents, dim=0).to(device)
+            else:
+                latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype).to(device)
+        else:
+            if latents.shape != shape:
+                raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {shape}")
+            latents = latents.to(device)
+
+        # scale the initial noise by the standard deviation required by the scheduler
+        latents = latents * self.scheduler.init_noise_sigma
+        return latents
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        # image: Union[PIL.Image.Image, List[PIL.Image.Image], torch.FloatTensor],
+        embeddings: torch.FloatTensor,
+        height: Optional[int] = None,
+        width: Optional[int] = None,
+        num_inference_steps: int = 50,
+        guidance_scale: float = 7.5,
+        num_images_per_prompt: Optional[int] = 1,
+        eta: float = 0.0,
+        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
+        latents: Optional[torch.FloatTensor] = None,
+        output_type: Optional[str] = "pil",
+        return_dict: bool = True,
+        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback_steps: Optional[int] = 1,
+    ):
+        r"""
+        Function invoked when calling the pipeline for generation.
+
+        Args:
+            image (`PIL.Image.Image` or `List[PIL.Image.Image]` or `torch.FloatTensor`):
+                The image or images to guide the image generation. If you provide a tensor, it needs to comply with the
+                configuration of
+                [this](https://huggingface.co/lambdalabs/sd-image-variations-diffusers/blob/main/feature_extractor/preprocessor_config.json)
+                `CLIPFeatureExtractor`
+            height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
+                The height in pixels of the generated image.
+            width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
+                The width in pixels of the generated image.
+            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.
+            guidance_scale (`float`, *optional*, defaults to 7.5):
+                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.
+            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`, *optional*):
+                One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
+                to make generation deterministic.
+            latents (`torch.FloatTensor`, *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`.
+            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.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
+                plain tuple.
+            callback (`Callable`, *optional*):
+                A function that will be called every `callback_steps` steps during inference. The function will be
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+            callback_steps (`int`, *optional*, defaults to 1):
+                The frequency at which the `callback` function will be called. If not specified, the callback will be
+                called at every step.
+
+        Returns:
+            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
+            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
+            When returning a tuple, the first element is a list with the generated images, and the second element is a
+            list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
+            (nsfw) content, according to the `safety_checker`.
+        """
+        # 0. Default height and width to unet
+        height = height or self.unet.config.sample_size * self.vae_scale_factor
+        width = width or self.unet.config.sample_size * self.vae_scale_factor
+
+        # 1. Check inputs. Raise error if not correct
+        # self.check_inputs(image, height, width, callback_steps)
+
+        # 2. Define call parameters
+        if isinstance(embeddings, list):
+            batch_size = len(embeddings)
+        else:
+            batch_size = 1
+        device = self._execution_device
+        # 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.
+        do_classifier_free_guidance = guidance_scale > 1.0
+
+        # 3. Encode input image
+        embeddings = self._prepare_embeddings(embeddings, device, num_images_per_prompt, do_classifier_free_guidance)
+        embeddings = embeddings.to(device)
+
+        # 4. Prepare timesteps
+        self.scheduler.set_timesteps(num_inference_steps, device=device)
+        timesteps = self.scheduler.timesteps
+
+        # 5. Prepare latent variables
+        num_channels_latents = self.unet.in_channels
+        latents = self.prepare_latents(
+            batch_size * num_images_per_prompt,
+            num_channels_latents,
+            height,
+            width,
+            embeddings.dtype,
+            device,
+            generator,
+            latents,
+        )
+
+        # 6. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
+        extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
+
+        # 7. Denoising loop
+        num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
+        with self.progress_bar(total=num_inference_steps) as progress_bar:
+            for i, t in enumerate(timesteps):
+                # expand the latents if we are doing classifier free guidance
+                latent_model_input = torch.cat([latents] * 2) if do_classifier_free_guidance else latents
+                latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
+
+                # predict the noise residual
+                noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=embeddings).sample
+
+                # perform guidance
+                if do_classifier_free_guidance:
+                    noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
+                    noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
+
+                # compute the previous noisy sample x_t -> x_t-1
+                latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs).prev_sample
+
+                # call the callback, if provided
+                if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
+                    progress_bar.update()
+                    if callback is not None and i % callback_steps == 0:
+                        callback(i, t, latents)
+
+        # 8. Post-processing
+        image = self.decode_latents(latents)
+
+        # 9. Run safety checker
+        image, has_nsfw_concept = self.run_safety_checker(image, device, embeddings.dtype)
+
+        # 10. Convert to PIL
+        if output_type == "pil":
+            image = self.numpy_to_pil(image)
+
+        if not return_dict:
+            return (image, has_nsfw_concept)
+
+        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

requirements.txt

@@ -0,0 +1,8 @@
+torch
+torchvision
+numpy
+transformers
+diffusers
+# ftfy
+gradio
+accelerate

test-platform.py

@@ -0,0 +1,8 @@
+import torch
+
+# set device to mps if avaliable, cude if avaliable, cpu otherwise
+device = torch.device("mps" if torch.backends.mps.is_available() else "cuda:0" if torch.cuda.is_available() else "cpu")
+print (device)
+
+x = torch.zeros(1, device=device)
+print (str(x))