From 9032fd56330dc024f707a5d7df24367fd0cdadf5 Mon Sep 17 00:00:00 2001 From: Haytham Abuelfutuh Date: Thu, 13 Jun 2024 13:27:12 -0700 Subject: [PATCH] Use currently installed unionai library versiong Signed-off-by: Haytham Abuelfutuh --- .../src/lip_sync_src/face3d/models/bfm.py | 2 +- .../lip_sync_src/face3d/models/facerecon_model.py | 14 +++++++------- .../src/lip_sync_src/generate_batch.py | 2 +- _blogs/video-translation/src/tasks/clone_voice.py | 3 --- .../src/tasks/fetch_audio_and_image.py | 2 -- _blogs/video-translation/src/tasks/lip_sync.py | 5 +---- _blogs/video-translation/src/tasks/speech2text.py | 3 --- .../video-translation/src/tasks/translate_text.py | 2 -- _blogs/video-translation/src/workflow.py | 10 +++++----- 9 files changed, 15 insertions(+), 28 deletions(-) diff --git a/_blogs/video-translation/src/lip_sync_src/face3d/models/bfm.py b/_blogs/video-translation/src/lip_sync_src/face3d/models/bfm.py index a4c03bbc..10702fd9 100644 --- a/_blogs/video-translation/src/lip_sync_src/face3d/models/bfm.py +++ b/_blogs/video-translation/src/lip_sync_src/face3d/models/bfm.py @@ -8,7 +8,7 @@ import torch.nn.functional as F from scipy.io import loadmat -from src.face3d.util.load_mats import transferBFM09 +from face3d.util.load_mats import transferBFM09 def perspective_projection(focal, center): diff --git a/_blogs/video-translation/src/lip_sync_src/face3d/models/facerecon_model.py b/_blogs/video-translation/src/lip_sync_src/face3d/models/facerecon_model.py index c900113e..9faab372 100644 --- a/_blogs/video-translation/src/lip_sync_src/face3d/models/facerecon_model.py +++ b/_blogs/video-translation/src/lip_sync_src/face3d/models/facerecon_model.py @@ -6,15 +6,15 @@ import trimesh from scipy.io import savemat -from src.face3d.models import networks -from src.face3d.models.base_model import BaseModel -from src.face3d.models.bfm import ParametricFaceModel -from src.face3d.models.losses import (landmark_loss, perceptual_loss, +from face3d.models import networks +from face3d.models.base_model import BaseModel +from face3d.models.bfm import ParametricFaceModel +from face3d.models.losses import (landmark_loss, perceptual_loss, photo_loss, reflectance_loss, reg_loss) -from src.face3d.util import util -from src.face3d.util.nvdiffrast import MeshRenderer +from face3d.util import util +from face3d.util.nvdiffrast import MeshRenderer -# from src.face3d.util.preprocess import estimate_norm_torch +# from face3d.util.preprocess import estimate_norm_torch class FaceReconModel(BaseModel): diff --git a/_blogs/video-translation/src/lip_sync_src/generate_batch.py b/_blogs/video-translation/src/lip_sync_src/generate_batch.py index b2028ed7..efc07098 100644 --- a/_blogs/video-translation/src/lip_sync_src/generate_batch.py +++ b/_blogs/video-translation/src/lip_sync_src/generate_batch.py @@ -6,7 +6,7 @@ import torch from tqdm import tqdm -import src.lip_sync_src.utils.audio as audio +import lip_sync_src.utils.audio as audio def crop_pad_audio(wav, audio_length): diff --git a/_blogs/video-translation/src/tasks/clone_voice.py b/_blogs/video-translation/src/tasks/clone_voice.py index 2885d3da..e4d673ba 100644 --- a/_blogs/video-translation/src/tasks/clone_voice.py +++ b/_blogs/video-translation/src/tasks/clone_voice.py @@ -27,11 +27,8 @@ clone_voice_image = ImageSpec( name="clone_voice", - builder="ucimage", packages=[ "TTS==0.22.0", - "flytekit==1.10.7", - "unionai==0.1.5", ], cuda="12.1.0", cudnn="8", diff --git a/_blogs/video-translation/src/tasks/fetch_audio_and_image.py b/_blogs/video-translation/src/tasks/fetch_audio_and_image.py index c41f7f5b..0063e917 100644 --- a/_blogs/video-translation/src/tasks/fetch_audio_and_image.py +++ b/_blogs/video-translation/src/tasks/fetch_audio_and_image.py @@ -9,12 +9,10 @@ preprocessing_image = ImageSpec( name="fetch_audio_and_image", - builder="ucimage", apt_packages=["ffmpeg"], packages=[ "moviepy==1.0.3", "katna==0.9.2", - "unionai==0.1.5", ], ) diff --git a/_blogs/video-translation/src/tasks/lip_sync.py b/_blogs/video-translation/src/tasks/lip_sync.py index 915e1aa7..e99bf725 100644 --- a/_blogs/video-translation/src/tasks/lip_sync.py +++ b/_blogs/video-translation/src/tasks/lip_sync.py @@ -10,7 +10,6 @@ lip_sync_image = ImageSpec( name="lip_sync", - builder="ucimage", apt_packages=["build-essential", "libssl-dev", "ffmpeg", "libsndfile1", "git"], packages=[ "setuptools==69.1.1", @@ -40,9 +39,7 @@ "av==11.0.0", "safetensors==0.4.2", "huggingface-hub==0.21.4", - "realesrgan==0.3.0", - "flytekit==1.10.7", - "unionai==0.1.5", + "realesrgan==0.3.0" ], cuda="12.1.0", cudnn="8", diff --git a/_blogs/video-translation/src/tasks/speech2text.py b/_blogs/video-translation/src/tasks/speech2text.py index f3080858..43322b5f 100644 --- a/_blogs/video-translation/src/tasks/speech2text.py +++ b/_blogs/video-translation/src/tasks/speech2text.py @@ -8,13 +8,10 @@ speech2text_image = ImageSpec( name="speech2text", - builder="ucimage", apt_packages=["ffmpeg"], packages=[ "transformers==4.36.2", "torch==2.2.1", - "flytekit==1.10.7", - "unionai==0.1.5", ], cuda="12.1.0", cudnn="8", diff --git a/_blogs/video-translation/src/tasks/translate_text.py b/_blogs/video-translation/src/tasks/translate_text.py index 07beedf5..a49cb951 100644 --- a/_blogs/video-translation/src/tasks/translate_text.py +++ b/_blogs/video-translation/src/tasks/translate_text.py @@ -107,7 +107,6 @@ language_translation_image = ImageSpec( name="language_translation", - builder="ucimage", packages=[ "transformers==4.36.2", "torch==2.2.1", @@ -116,7 +115,6 @@ "flytekit==1.10.7", "sentencepiece==0.2.0", "nltk==3.8.1", - "unionai==0.1.5", ], ) diff --git a/_blogs/video-translation/src/workflow.py b/_blogs/video-translation/src/workflow.py index 623b0731..f44be23e 100644 --- a/_blogs/video-translation/src/workflow.py +++ b/_blogs/video-translation/src/workflow.py @@ -3,11 +3,11 @@ from flytekit import workflow from flytekit.types.file import FlyteFile -from .tasks.clone_voice import clone_voice -from .tasks.fetch_audio_and_image import fetch_audio_and_image -from .tasks.lip_sync import lip_sync -from .tasks.speech2text import speech2text -from .tasks.translate_text import translate_text +from tasks.clone_voice import clone_voice +from tasks.fetch_audio_and_image import fetch_audio_and_image +from tasks.lip_sync import lip_sync +from tasks.speech2text import speech2text +from tasks.translate_text import translate_text @workflow