Vision & Multimodal

ViT, CLIP, diffusion, SAM, and the vision-language models that read images as tokens.

12concepts
131flashcards
101minutes of reading
  1. 01 Contrastive Vision-Language: CLIP How a 400M image-text contrastive objective produced a shared embedding space that does zero-shot classification, retrieval, and grounding without any task-specific labels. intermediate 8m 5 cards
  2. 02 Multimodal Hallucination and Benchmark Validity Why vision-language models describe objects that are not in the image, and why a large share of multimodal benchmark scores can be reproduced without showing the model any image at all. intermediate 8m 18 cards
  3. 03 Native-Resolution Vision Encoding Why squashing every image to a fixed square destroyed text legibility in early VLMs, and how tiling, patch packing and dynamic resolution replaced it at the cost of an unbounded visual token budget. intermediate 9m 18 cards
  4. 04 Segment Anything (SAM) and Dense Prediction How promptable segmentation became a foundation-model task, what SAM's encoder-decoder split was designed for, and where it still loses to specialist models. intermediate 7m 5 cards
  5. 05 Video, Audio, and Any-to-Any Models How Whisper, V-JEPA, Sora-class video generators, MusicGen, and unified any-to-any models extend the multimodal stack beyond static images. intermediate 9m 6 cards
  6. 06 Vision Transformers (ViT) How treating an image as a sequence of patches let pure transformers beat CNNs once data crossed the 300M-image mark, and what the architecture gave up to get there. intermediate 8m 5 cards
  7. 07 Visual Document Retrieval Why the OCR-parse-chunk-embed pipeline loses exactly the documents that matter, and how late-interaction retrieval over page images replaces five brittle stages with one model. intermediate 8m 15 cards
  8. 08 Diffusion Models How learning to invert a noise process became the dominant generative recipe for images, video, and audio, and why Flow Matching and DiTs are reshaping the recipe in 2024. advanced 10m 6 cards
  9. 09 Discrete Visual Tokenisers How VQ-VAE, VQGAN and FSQ turn an image into a short sequence of integers, why codebook collapse wrecks half of them, and what the compression ratio costs in reconstruction fidelity. advanced 8m 15 cards
  10. 10 Early-Fusion Mixed-Modal Models The difference between bolting a vision encoder onto a frozen LLM and training one transformer over interleaved image and text tokens from scratch, and why the second approach destabilises training. advanced 8m 15 cards
  11. 11 Flow Matching and Rectified Flow Why regressing a velocity field along straight noise-to-data paths replaced score matching in frontier image models, and what the straight-line claim does and does not guarantee. advanced 9m 18 cards
  12. 12 Multimodal LLMs: LLaVA, Flamingo, GPT-4V The vision-encoder-plus-projector-plus-LLM recipe that dominates open multimodal models, why Flamingo's perceiver design still matters for video, and what native-multimodal frontier models do differently. advanced 9m 5 cards