I build WebGPU PyTorch debugger, dataset browser and NumPy viewer

Svetovid, a new WebGPU-powered tool, enables real-time visualization and debugging of PyTorch models, NumPy arrays, and large datasets directly in the browser. The tool supports interactive exploration of neural network architectures, activation inspection, and streaming of HuggingFace datasets, with examples showing integration with GPT-2, DETR, and convolutional layers.

Get empowered by Svetovid , whose four heads represent omniscience: the ability to see through space and time in all directions at once. Maintain unified perspective across your tensors, models, and massive datasets. Peer into the architecture of your neural networks and navigate high-dimensional data with absolute clarity. array inspector.py NumPy Inspector Visualize any NumPy array with instant dimension awareness. Support for high-dimensional, large arrays. python import svetoviz webgpu as sv data = np.random.rand 4, 4, 32, 32 .astype np.float32 sv.array data, gap size=10 Learn More content/guides/numpy.html pytorch debugger.py PyTorch Debugger Directly load your PyTorch module, interact, and debug activations in real-time. python import svetoviz webgpu as sv img = Image.open "sample image.jpg" .convert "RGB" img np = np.array img .astype np.float32 / 255.0 Rearrange from H, W, C to C, H, W and add batch dim input tensor = torch.from numpy img np .permute 2, 0, 1 .unsqueeze 0 2. Define the 2D Convolutional module conv2d = nn.Conv2d in channels=3, out channels=16, kernel size=3, stride=1, padding=1 def terminal callback buffer, message, images, files : 3. Process the image through the module output = conv2d input tensor Log the feature map shape to the Svetoviz terminal buffer.send system message f"Processed image. Shape: {list output.shape }" 4. Start the interactive session sv.pytorch module=conv2d, terminal callback=terminal callback Learn More content/guides/pytorch.html model viewer.py Model Analysis Deep-dive into model architecture. Explore layers down to a single parameter. python import svetoviz webgpu as sv image processor = DetrImageProcessor.from pretrained "facebook/detr-resnet-50", revision="no timm", device map="cpu" model = DetrForObjectDetection.from pretrained "facebook/detr-resnet-50", revision="no timm", device map="cpu" sv.model model=model, image processor=image processor Learn More content/guides/models.html model debug.py Model Debugger Interact with your model and investigate activations in real-time. python import svetoviz webgpu as sv tokenizer = GPT2Tokenizer.from pretrained 'gpt2', device map="cpu" model = GPT2LMHeadModel.from pretrained 'gpt2', device map="cpu" def terminal callback buffer, message, images, files : text = message "Replace me by any text you'd like." buffer.send user message text inputs = tokenizer text, return tensors="pt" Generate continuation outputs = model.generate inputs, max new tokens=50 Decode to string decoded = tokenizer.decode outputs 0 , skip special tokens=True buffer.send system message decoded sv.model model=model, tokenizer=tokenizer, terminal callback=terminal callback Learn More content/guides/models.html dataset handler.py Dataset Explorer Efficiently stream and visualize large-scale HuggingFace datasets from local storage or remote server. python import svetoviz webgpu as sv hf name = "eltorio/ROCOv2-radiology" dataset = load dataset hf name, split="train" view = DatasetView name=hf name, dataset=dataset, image columns= "image" , cell width=400, cell height=400 sv.save to disc view=view, compression="png 0", directory=f"/Volumes/Untitled/{hf name}" sv.load from disc directory=f"/Volumes/Untitled/{hf name}" Learn More content/guides/datasets.html image browser.py Image Browser Browse through massive collections of images regardless of resolution. python import svetoviz webgpu as sv view = DirectorySpiralView directory="/Desktop/nasaimages" sv.save to disc view=view, compression="jpeg 0", directory="/Volumes/Untitled/universe" sv.load from disc directory="/Volumes/Untitled/universe" Learn More content/guides/images.html Installation