Run SVD locally

svd.py

# from cog import BasePredictor, Input, Path
import os
import cv2
import time
import math
import torch
import numpy as np
from PIL import Image
from glob import glob
from typing import Optional
from omegaconf import OmegaConf
from einops import rearrange, repeat
from torchvision.transforms import ToTensor
from sgm.inference.helpers import embed_watermark
from sgm.util import default, instantiate_from_config

"""Exported from stability/ai generative-models """
def get_unique_embedder_keys_from_conditioner(conditioner):
    return list(set([x.input_key for x in conditioner.embedders]))

def get_batch(keys, value_dict, N, T, device):
    batch = {}
    batch_uc = {}

    for key in keys:
        if key == "fps_id":
            batch[key] = (
                torch.tensor([value_dict["fps_id"]])
                .to(device)
                .repeat(int(math.prod(N)))
            )
        elif key == "motion_bucket_id":
            batch[key] = (
                torch.tensor([value_dict["motion_bucket_id"]])
                .to(device)
                .repeat(int(math.prod(N)))
            )
        elif key == "cond_aug":
            batch[key] = repeat(
                torch.tensor([value_dict["cond_aug"]]).to(device),
                "1 -> b",
                b=math.prod(N),
            )
        elif key == "cond_frames":
            batch[key] = repeat(value_dict["cond_frames"], "1 ... -> b ...", b=N[0])
        elif key == "cond_frames_without_noise":
            batch[key] = repeat(
                value_dict["cond_frames_without_noise"], "1 ... -> b ...", b=N[0]
            )
        else:
            batch[key] = value_dict[key]

    if T is not None:
        batch["num_video_frames"] = T

    for key in batch.keys():
        if key not in batch_uc and isinstance(batch[key], torch.Tensor):
            batch_uc[key] = torch.clone(batch[key])
    return batch, batch_uc

def load_model(
    config: str,
    device: str,
    num_frames: int,
    num_steps: int,
):
    config = OmegaConf.load(config)
    if device == "cuda":
        config.model.params.conditioner_config.params.emb_models[
            0
        ].params.open_clip_embedding_config.params.init_device = device

    config.model.params.sampler_config.params.num_steps = num_steps
    config.model.params.sampler_config.params.guider_config.params.num_frames = (
        num_frames
    )
    if device == "cuda":
        with torch.device(device):
            model = instantiate_from_config(config.model).to(device).eval()
    else:
        model = instantiate_from_config(config.model).to(device).eval()

    return model



def predict(
    input_image
) -> str:
    """Run a single prediction on the model"""
    num_frames: Optional[int] = 14
    num_steps: Optional[int] = 25
    fps_id: int = 6
    motion_bucket_id: int = 127
    cond_aug: float = 0.02
    seed: int = 23
    decoding_t: int = 7
    device: str = "cuda"
    output_folder: Optional[str] = "output/"
    model_config = "svd.yaml"

    model = load_model(
        model_config,
        device,
        num_frames,
        num_steps,
    )
    torch.manual_seed(seed)

    input_img_path = input_image
    output_path = None
    with Image.open(input_img_path) as image:
        if image.mode == "RGBA":
            image = image.convert("RGB")
        w, h = image.size

        if h % 64 != 0 or w % 64 != 0:
            width, height = map(lambda x: x - x % 64, (w, h))
            image = image.resize((width, height))
            print(
                f"WARNING: Your image is of size {h}x{w} which is not divisible by 64. We are resizing to {height}x{width}!"
            )

        image = ToTensor()(image)
        image = image * 2.0 - 1.0

    image = image.unsqueeze(0).to(device)
    H, W = image.shape[2:]
    assert image.shape[1] == 3
    F = 8
    C = 4
    shape = (num_frames, C, H // F, W // F)
    if (H, W) != (576, 1024):
        print(
            "WARNING: The conditioning frame you provided is not 576x1024. This leads to suboptimal performance as model was only trained on 576x1024. Consider increasing `cond_aug`."
        )
    if motion_bucket_id > 255:
        print(
            "WARNING: High motion bucket! This may lead to suboptimal performance."
        )

    if fps_id < 5:
        print("WARNING: Small fps value! This may lead to suboptimal performance.")

    if fps_id > 30:
        print("WARNING: Large fps value! This may lead to suboptimal performance.")

    value_dict = {}
    value_dict["motion_bucket_id"] = motion_bucket_id
    value_dict["fps_id"] = fps_id
    value_dict["cond_aug"] = cond_aug
    value_dict["cond_frames_without_noise"] = image
    value_dict["cond_frames"] = image + cond_aug * torch.randn_like(image)
    value_dict["cond_aug"] = cond_aug

    with torch.no_grad():
        with torch.autocast(device):
            batch, batch_uc = get_batch(
                get_unique_embedder_keys_from_conditioner(model.conditioner),
                value_dict,
                [1, num_frames],
                T=num_frames,
                device=device,
            )
            c, uc = model.conditioner.get_unconditional_conditioning(
                batch,
                batch_uc=batch_uc,
                force_uc_zero_embeddings=[
                    "cond_frames",
                    "cond_frames_without_noise",
                ],
            )

            for k in ["crossattn", "concat"]:
                uc[k] = repeat(uc[k], "b ... -> b t ...", t=num_frames)
                uc[k] = rearrange(uc[k], "b t ... -> (b t) ...", t=num_frames)
                c[k] = repeat(c[k], "b ... -> b t ...", t=num_frames)
                c[k] = rearrange(c[k], "b t ... -> (b t) ...", t=num_frames)

            randn = torch.randn(shape, device=device)

            additional_model_inputs = {}
            additional_model_inputs["image_only_indicator"] = torch.zeros(
                2, num_frames
            ).to(device)
            additional_model_inputs["num_video_frames"] = batch["num_video_frames"]

            def denoiser(input, sigma, c):
                return model.denoiser(
                    model.model, input, sigma, c, **additional_model_inputs
                )

            samples_z = model.sampler(denoiser, randn, cond=c, uc=uc)
            model.en_and_decode_n_samples_a_time = decoding_t
            samples_x = model.decode_first_stage(samples_z)
            samples = torch.clamp((samples_x + 1.0) / 2.0, min=0.0, max=1.0)

            os.makedirs(output_folder, exist_ok=True)
            base_count = len(glob(os.path.join(output_folder, "*.mp4")))
            video_path = os.path.join(output_folder, f"{base_count:06d}.mp4")
            writer = cv2.VideoWriter(
                video_path,
                cv2.VideoWriter_fourcc(*"MP4V"),
                fps_id + 1,
                (samples.shape[-1], samples.shape[-2]),
            )
            output_path = video_path

            samples = embed_watermark(samples)
            vid = (
                (rearrange(samples, "t c h w -> t h w c") * 255)
                .cpu()
                .numpy()
                .astype(np.uint8)
            )
            for frame in vid:
                frame = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
                writer.write(frame)
            writer.release()

    return output_path

t1 = time.time()
for _ in range(10):
    predict("fireworks.png")
t2 = time.time()
print("Inference took - ",t2-t1,"seconds")