💡 Lumina-T2X: Transforming Text into Any Modality, Resolution, and Duration via Flow-based Large Diffusion Transformers


▪Github: https://github.com/alpha-vllm/lumina-t2x
▪Paper: https://arxiv.org/abs/2405.05945
▪Demo: https://lumina.sylin.host/

@Machine_learn

⚡️ Deblur-GS: 3D Gaussian Splatting from Camera Motion Blurred Images


▪Code: https://github.com/Chaphlagical/Deblur-GS
▪Paper: https://chaphlagical.icu/Deblur-GS/static/paper/Deblur_GS_author_version.pdf
▪Project: https://chaphlagical.icu/Deblur-GS/

@Machine_learn

Images that Sound: Composing Images and Sounds on a Single Canvas

abs: https://arxiv.org/abs/2405.12221
project page: https://ificl.github.io/images-that-sound/
code: https://github.com/IFICL/images-that-sound

This paper introduces an inference-time procedure that generates images that are also spectrograms corresponding to the prompt. It uses a latent image and audio diffusion model with same latent space (Stable Diffusion v1.5 and Auffusion) and denoise the same latent with both.

@Machine_learn

🔥🔥🔥 YOLOv10: Real-Time End-to-End Object Detection

▪Paper: arxiv.org/pdf/2405.14458
▪Github: https://github.com/THU-MIG/yolov10/
▪Demo :https://huggingface.co/spaces/kadirnar/Yolov10
▪Colab: https://colab.research.google.com/github/roboflow-ai/notebooks/blob/main/notebooks/train-yolov10-object-detection-on-custom-dataset.ipynb#scrollTo=SaKTSzSWnG7s

@machine_learn

با عرض سلام این مقاله رو می خواییم برای Nature بفرستیم جایگاه های ۱ تا ۴ اش خالیه از دوستان کسی نیاز داشت در خدمتیم
Title:
Detection of brain tumors from images using the UNet architecture, with a comparative analysis of transfer learning methods and CNNs.
——————————————————————--
Abstract:
Health is crucial for human life, especially brain health, which is vital for all executive functions. Diagnosing brain health issues is often done using magnetic resonance imaging (MRI) devices, which provide critical data for health decision-makers. Images from these devices serve as a significant source of big data for artificial intelligence applications. This big data facilitates high performance in image processing classification problems, a subfield of artificial intelligence. In this study, we aim to classify brain tumors such as glioma, meningioma, and pituitary tumors from brain MRI images using the UNet architecture. To compare the results and gain a better understanding, we also employed Convolutional Neural Networks (CNN) and CNN-based models like Inception-V3, EfficientNetB4, VGG19, along with transfer learning methods for classification tasks. The models were evaluated using F-score, recall, precision, and accuracy metrics. The best accuracy result was achieved with CNN-VGG16, reaching 97%. The same transfer learning model also showed an F-score of 96%, an Area Under the Curve (AUC) value of 98%, a recall value of 98%, and a precision value of 97%. The UNet architecture and CNN-based transfer learning models play a significant role in the early diagnosis and rapid treatment of brain tumors, which is vital for improving patient outcomes.
——————————————————————
Keywords:
Brain tumor detection, UNet, CNN, Transfer Learning.
——————————————————————
Journal:
Scientific Reports

@Raminmousa
@Machine_learn
@paper4money

سلام دوستان حداقل ماين مي كنينن NFT ماين كنين كه يه چيزي گيرتون بياد. به نظرم اساس كوين هارو بخونين بعد ماين كنين. پروژه پايين از تمامي مواردي كه فرستادين برام بهتر بوده.

https://www.tg-me.com/SpinnerCoin_bot/app?startapp=r_280673

🎙 Real-time in-browser speech recognition



▪Сode: https://github.com/xenova/transformers.js/tree/v3/examples/webgpu-whisper
▪Hf: https://huggingface.co/spaces/Xenova/realtime-whisper-webgpu

@Machine_learn

https://www.tg-me.com/SpinnerCoin_bot/app?startapp=r_280673


تنها پروژه اي كه اين روزا از نظرم اهميت داره Spinner هستش كه به ماين NFT مي پردازه.
💎 +750 $SPN as a first-time gift

🔈 Separating the "Chirp" from the "Chat": Self-supervised Visual Grounding of Sound and Language



▪Paper: https://arxiv.org/abs/2406.05629
▪Website: https://mhamilton.net/denseav
▪Code: https://github.com/mhamilton723/DenseAV
▪Video: https://youtu.be/wrsxsKG-4eE

@Machine_learn

🔥 Astrologers have announced a week of video generation models!

Following the hype around the Kling, Luma and Runway models, a new open source version of Open-Sora has been released.

Open-Sora 1.2 from Hpcoretech has been published on huggingface.

Basic moments:

The new 1.1B model is trained on 20M videos and generates videos up to 14 seconds long at 720p resolution.

▪Diffusion Model: https://huggingface.co/hpcai-tech/OpenSora-STDiT-v3
▪VAE model: https://huggingface.co/hpcai-tech/OpenSora-VAE-v1.2
▪Technical report: https://github.com/hpcaitech/Open-Sora/blob/main/docs/report_03.md
▪Demo: https://huggingface.co/spaces/hpcai-tech/open-sora

@Machine_learn