Text2Mesh is a framework to stylize meshes by predicting color and geometric details which conform to a target text prompt. Optimization occurs through rendered 2D images & augmentations, then using CLIP similarity to train the network to produce consistent stylized meshes.
2/6
The technique leverages the representational power of CLIP and is capable of generating a variety of different local geometric displacements to synthesize a wide range of styles.
Deep learning models for tabular data continue to improve. What are the latest methods and recent progress?
Let’s have a look ↓
1) Wide&Deep jointly trains wide linear models and deep neural networks to combine the benefits of memorization and generalization for real-world recommender systems. The model was productionized and evaluated on Google Play.
2) TaBERT is a pretrained LM that jointly learns representations for natural language sentences and (semi-)structured tables. TaBERT works well for semantic parsing and is trained on a large corpus of 26 million tables and their English contexts.
Graph neural networks are driving lots of progress in machine learning by extending deep learning approaches to complex graph data and applications.
Let’s take a look at a few methods ↓
1) A Graph Convolutional Network, or GCN, is an approach for semi-supervised learning on graph-structured data. It’s based on an efficient variant of CNNs which operates directly on graphs and is useful for semi-supervised node classification.
2) Diffusion-convolutional neural networks (DCNN) introduce a diffusion-convolution operation to extend CNNs to graph data. This enables learning of diffusion-based representations. It's used as an effective basis for node classification.
It proposes architectural changes that suppress aliasing and forces the model to implement more natural hierarchical refinement which improves its ability to generate video and animation.
In the cinemagraph below, we can see that in StyleGAN2 the texture (e.g., wrinkles and hairs) appears to stick to the screen coordinates. In comparison, StyleGAN3 (right) transforms details coherently:
2/8
The following example shows the same issue with StyleGAN2: textural details appear fixed. As for alias-free StyleGAN3, smooth transformations with the rest of the screen can be seen.
In a new paper from @wightmanr et al. a traditional ResNet-50 is re-trained using a modern training protocol. It achieves a very competitive 80.4% top-1 accuracy on ImageNet without using extra data or distillation.
The paper catalogues the exact training settings to provide a robust baseline for future experiments:
It also records training costs and inference times on ImageNet classification between other architectures trained with the proposed ResNet-50 optimized training procedure:
🚨 Newsletter Issue #3. Featuring a new state-of-the-art on ImageNet, a trillion-parameter language model, 10 applications of transformers you didn’t know about, and much more! Read on below: