Add blog link and update status

Abhinay1997 · Oct 4, 2024 · d56330e · d56330e
1 parent dea7ebd
commit d56330e
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/README.md b/README.md
@@ -3,13 +3,15 @@
 | Idea                                               | Status                                                                                     | References/Papers                                                 | 
 |----------------------------------------------------|------------------------------------------------------------------------------------------------------|---------------------------------------------------------|
 | FIFO CogVideoX                                     | In Progress                                                                                |   https://jjihwan.github.io/projects/FIFO-Diffusion |
+|Comparision of different cfg like methods| In Progress                                   |  Smoothed Energy Guidance, Guidance embedding, CFG and CFG++| 
 | Flux Image Inversion using RNRI                    | Blocked. Understanding of prior for flow matching too low                                  |   https://barakmam.github.io/rnri.github.io/| 
 | CogVideoX Attention Scaling                        | Paused. Need to recheck for higher res                                                     |   https://arxiv.org/abs/2306.08645| 
 | RB Modulation for FLUX                             | Planned. SOC is straightforward. AFA needs to be seen in detail. Planning to replace CSD with other image opeerators for different manifold explorations. Soft histograms for relative color palette retention ?                            |   https://rb-modulation.github.io/| 
 | CogVideoX distillation using FineVideo and PeRFlow | Planned. Needs compute grant. May be scrapped once BFLs video model is out.                |   https://arxiv.org/abs/2405.07510| 
 | Underwater Image Colorization as an Inverse Problem| Planned. Needs better underestanding of inverse problems                                   |   https://github.com/LituRout/PSLD| 
-|Comparision of different cfg like methods| Planned.                                   |  Smoothed Energy Guidance, Guidance embedding, CFG and CFG++| 
 | Flux generation steering using SAE for CLIP        | Planned. Need better understanding of SAEs & apply them to T5 as well |   https://www.lesswrong.com/posts/Quqekpvx8BGMMcaem/interpreting-and-steering-features-in-images| 
 | LoRA (move to MoRA ?) ControlNet layer        | Planned. Compute ∆W for Flux dev & its controlnet layer. Decompose to LoRA and see decomposition error. If its low enough, LoRA should be enough |   [ChatGPT conversation](https://chatgpt.com/share/66f12970-6608-800f-a24e-20c4d2766c4a)| 
 | MoRA finetuning Flux              | I have a hypothesis: MoRA might give better samples than LoRA for Flux. I'll try it out sometime next week maybe. TLDR: 1. Full finetuning > LoRA for personalization. 2. Full finetuning > MoRA > DoRA > LoRA. 3. MoRA should converge fast like LoRA but give better quality/diversity like finetuning. There should be no free lunch though. Hmm | 1. [MoRA: High-rank PEFT Approach](https://linkedin.com/pulse/mora-high-rank-peft-approach-fine-tuning-himank-jain-ewe3f/) 2. [Full Finetuning of Flux](https://dev.to/furkangozukara/full-fine-tuning-of-flux-yields-way-better-results-than-lora-training-as-expected-overfitting-and-bleeding-reduced-a-lot-3g6g) 3. [GitHub: MoRA](https://github.com/kongds/MoRA) |
 | Transformer layers as Painters for DiTs            | Complete. Results published [here](https://huggingface.co/blog/NagaSaiAbhinay/transformer-layers-as-painters-dit)|   https://arxiv.org/abs/2407.09298| 
+
+Checkout my notes/blog here: [abhinay1997.github.io](https://abhinay1997.github.io)