Jaward Sesay
@jawardsesay.bsky.social
Building Lectūra AI | Chief nanoAI Officer:) | 1st Paper (AutoAgents) Accepted @ IJCAI 2024 | Open-sourcerer | Sensei: Karpathy
https://github.com/Jaykef
https://github.com/Jaykef
Lightweight implementation of a Rectified Flow Transformer model, ~ 618k parameters, 6 layers deep, dim 64, patch size 4, learning rate 5e-4 trained on my 8bg ram m2 macbookair for 2k epochs. #AI #LLM #Diffusion
June 27, 2025 at 7:37 AM
Lightweight implementation of a Rectified Flow Transformer model, ~ 618k parameters, 6 layers deep, dim 64, patch size 4, learning rate 5e-4 trained on my 8bg ram m2 macbookair for 2k epochs. #AI #LLM #Diffusion
Lightweight implementation of the seminal paper “Sequence to Sequence Learning with Neural Networks”
Built, trained and eval a 2 layer deep seq2seq LSTM-based model (~10M params) on German-English corpus of Multi30K dataset. In honor of Ilya for winning this year’s NeurIPs Test of Time paper award.
Built, trained and eval a 2 layer deep seq2seq LSTM-based model (~10M params) on German-English corpus of Multi30K dataset. In honor of Ilya for winning this year’s NeurIPs Test of Time paper award.
December 9, 2024 at 12:16 AM
Lightweight implementation of the seminal paper “Sequence to Sequence Learning with Neural Networks”
Built, trained and eval a 2 layer deep seq2seq LSTM-based model (~10M params) on German-English corpus of Multi30K dataset. In honor of Ilya for winning this year’s NeurIPs Test of Time paper award.
Built, trained and eval a 2 layer deep seq2seq LSTM-based model (~10M params) on German-English corpus of Multi30K dataset. In honor of Ilya for winning this year’s NeurIPs Test of Time paper award.
Implements compute-efficient DeepPCR algorithm which parallelizes sequential operations thus speeding up inference and training of neural networks.
DeepPCR can significantly reduce the time complexity in operations such as denoising in latent diffusion from O(L) to O(log2 L).
DeepPCR can significantly reduce the time complexity in operations such as denoising in latent diffusion from O(L) to O(log2 L).
December 1, 2024 at 1:33 AM
Implements compute-efficient DeepPCR algorithm which parallelizes sequential operations thus speeding up inference and training of neural networks.
DeepPCR can significantly reduce the time complexity in operations such as denoising in latent diffusion from O(L) to O(log2 L).
DeepPCR can significantly reduce the time complexity in operations such as denoising in latent diffusion from O(L) to O(log2 L).
If you’re interested in anything AI (especially research) let’s linkup:)
November 24, 2024 at 5:04 AM
If you’re interested in anything AI (especially research) let’s linkup:)