Francesco215 / autoregressive_diffusionLinks
Video Diffusion Model. Autoregressive, long context, efficient training and inference. WIP
☆34Updated last month
Alternatives and similar repositories for autoregressive_diffusion
Users that are interested in autoregressive_diffusion are comparing it to the libraries listed below
Sorting:
- ☆102Updated 2 months ago
- WIP☆93Updated last year
- Getting crystal-like representations with harmonic loss☆193Updated 6 months ago
- $100K or 100 Days: Trade-offs when Pre-Training with Academic Resources☆147Updated 2 weeks ago
- ☆34Updated last year
- Focused on fast experimentation and simplicity☆75Updated 9 months ago
- The boundary of neural network trainability is fractal☆217Updated last year
- RS-IMLE☆42Updated 10 months ago
- ☆33Updated 9 months ago
- ☆52Updated last year
- σ-GPT: A New Approach to Autoregressive Models☆68Updated last year
- Code for the Fractured Entangled Representation Hypothesis position paper!☆200Updated 5 months ago
- Code to reproduce "Transformers Can Do Arithmetic with the Right Embeddings", McLeish et al (NeurIPS 2024)☆192Updated last year
- 📄Small Batch Size Training for Language Models☆63Updated 2 weeks ago
- [ICML 2025] Roll the dice & look before you leap: Going beyond the creative limits of next-token prediction☆72Updated 4 months ago
- ☆81Updated last year
- NeuMeta transforms neural networks by allowing a single model to adapt on the fly to different sizes, generating the right weights when n…☆43Updated 11 months ago
- ☆56Updated 10 months ago
- Implementation of Gradient Agreement Filtering, from Chaubard et al. of Stanford, but for single machine microbatches, in Pytorch☆25Updated 8 months ago
- Synthetic Alphabet Dataset☆19Updated 6 months ago
- DeMo: Decoupled Momentum Optimization☆194Updated 10 months ago
- Explorations into the proposal from the paper "Grokfast, Accelerated Grokking by Amplifying Slow Gradients"☆102Updated 9 months ago
- The official github repo for "Diffusion Language Models are Super Data Learners".☆134Updated 2 weeks ago
- LLMs represent numbers on a helix and manipulate that helix to do addition.☆26Updated 8 months ago
- Don't just regulate gradients like in Muon, regulate the weights too☆28Updated 2 months ago
- The official repository for HyperZ⋅Z⋅W Operator Connects Slow-Fast Networks for Full Context Interaction.☆39Updated 6 months ago
- Latent Program Network (from the "Searching Latent Program Spaces" paper)☆98Updated 2 weeks ago
- Implementation of the proposed Spline-Based Transformer from Disney Research☆104Updated 11 months ago
- Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training☆132Updated last year
- Landing repository for the paper "Softpick: No Attention Sink, No Massive Activations with Rectified Softmax"☆85Updated last month