BlinkDL / SmallInitEmbLinks
LayerNorm(SmallInit(Embedding)) in a Transformer to improve convergence
☆58Updated 3 years ago
Alternatives and similar repositories for SmallInitEmb
Users that are interested in SmallInitEmb are comparing it to the libraries listed below
Sorting:
- Implementation of GateLoop Transformer in Pytorch and Jax☆90Updated last year
- Engineering the state of RNN language models (Mamba, RWKV, etc.)☆32Updated last year
- Yet another random morning idea to be quickly tried and architecture shared if it works; to allow the transformer to pause for any amount…☆52Updated 2 years ago
- HomebrewNLP in JAX flavour for maintable TPU-Training☆51Updated last year
- Exploring an idea where one forgets about efficiency and carries out attention across each edge of the nodes (tokens)☆55Updated 6 months ago
- ☆31Updated last year
- My explorations into editing the knowledge and memories of an attention network☆34Updated 2 years ago
- Automatically take good care of your preemptible TPUs☆37Updated 2 years ago
- ResiDual: Transformer with Dual Residual Connections, https://arxiv.org/abs/2304.14802☆96Updated 2 years ago
- Explorations into the recently proposed Taylor Series Linear Attention☆99Updated last year
- ☆81Updated last year
- Latent Diffusion Language Models☆68Updated 2 years ago
- RWKV model implementation☆38Updated 2 years ago
- Griffin MQA + Hawk Linear RNN Hybrid☆89Updated last year
- ☆34Updated last year
- ☆53Updated last year
- Exploration into the proposed "Self Reasoning Tokens" by Felipe Bonetto☆57Updated last year
- ☆19Updated 5 months ago
- Some personal experiments around routing tokens to different autoregressive attention, akin to mixture-of-experts☆119Updated last year
- Code for the paper "The Impact of Positional Encoding on Length Generalization in Transformers", NeurIPS 2023☆136Updated last year
- Implementation of Gated State Spaces, from the paper "Long Range Language Modeling via Gated State Spaces", in Pytorch☆101Updated 2 years ago
- Official Repository of Pretraining Without Attention (BiGS), BiGS is the first model to achieve BERT-level transfer learning on the GLUE …☆114Updated last year
- Official repository for the paper "Approximating Two-Layer Feedforward Networks for Efficient Transformers"☆38Updated 4 months ago
- Unofficial but Efficient Implementation of "Mamba: Linear-Time Sequence Modeling with Selective State Spaces" in JAX☆88Updated last year
- Fast, Modern, and Low Precision PyTorch Optimizers☆116Updated last month
- Training and evaluation code for the paper "Headless Language Models: Learning without Predicting with Contrastive Weight Tying" (https:/…☆27Updated last year
- ☆21Updated 11 months ago
- some common Huggingface transformers in maximal update parametrization (µP)☆86Updated 3 years ago
- Implementation of Gradient Agreement Filtering, from Chaubard et al. of Stanford, but for single machine microbatches, in Pytorch☆25Updated 9 months ago
- Explorations into the proposal from the paper "Grokfast, Accelerated Grokking by Amplifying Slow Gradients"☆102Updated 10 months ago