Hugging Face's logo
Harsha Vardhan Khurdula's picture
1 4 12

Harsha Vardhan Khurdula

HV-Khurdula
·

AI & ML interests

Transformers, Vision-Transformers, Multi-Agent Frameworks, Datatset Curation & Benchmarking.

Recent Activity

updated a dataset 10 days ago
JigsawStack/fleurs_mlt_benchmark
published a dataset 10 days ago
JigsawStack/fleurs_mlt_benchmark
reacted to Kseniase's post with 🚀 17 days ago
13 New types of LoRA LoRA (Low-Rank Adaptation) is a popular lightweight method for fine-tuning AI models. It doesn't update the full model, it adds small trainable components, low-rank matrices, while keeping the original weights frozen. Only these adapters are trained. Recently, many interesting new LoRA variations came out, so it’s a great time to take a look at these 13 clever approaches: 1. T-LoRA → https://huggingface.co/papers/2507.05964 A timestep-dependent LoRA method for adapting diffusion models with a single image. It dynamically adjusts updates and uses orthogonal initialization to reduce overlap, achieving better fidelity–alignment balance than standard LoRA 2. SingLoRA → https://huggingface.co/papers/2507.05566 Simplifies LoRA by using only one small matrix instead of usual two, and multiplying it by its own transpose (like A × Aᵀ). It uses half the parameters of LoRA and avoids scale mismatch between different matrices 3. LiON-LoRA → https://huggingface.co/papers/2507.05678 Improves control and precision in video diffusion models when training data is limited. It builds on LoRA, adding 3 key principles: linear scalability, orthogonality, and norm consistency. A controllable token and modified self-attention enables smooth adjustment of motion 4. LoRA-Mixer → https://huggingface.co/papers/2507.00029 Combines LoRA and mixture-of-experts (MoE) to adapt LLMs for multiple tasks. It dynamically routes task-specific LoRA experts into linear projections of attention modules, supporting both joint training and frozen expert reuse 5. QR-LoRA → https://huggingface.co/papers/2507.04599 Separates content and style when combining multiple LoRA adapters. It implements QR decomposition to structure parameter updates, where the orthogonal Q matrix reduces interference between features, and the R matrix captures specific transformations Read further in the comments 👇 If you like it, also subscribe to the Turing Post: https://www.turingpost.com/subscribe
View all activity

Organizations

Purdue University's profile picture JigsawStack's profile picture