Upload folder using huggingface_hub

README.md

@@ -3,7 +3,108 @@ license: apache-2.0
 base_model:
 - Lightricks/LTX-2.3
 tags:
-- inpaint
 - ltx
 - lora
----
+- inpaint
+---
+
+# LoRAs for LTX 2.3
+
+Here I will share some LoRAs that I trained for **LTX 2.3**.
+
+These LoRAs may cover different use cases over time, so this repository is not limited to inpainting only.
+
+## Models
+
+| File | Description |
+|---|---|
+| `ltx23_inpaint_rank128_v1_02500steps.safetensors` | Follows the prompt better, probably because it experienced less overfitting. |
+| `ltx23_inpaint_rank128_v1_10000steps.safetensors` | Follows the prompt in a more limited way, but uses the mask area better. This is probably because it experienced more overfitting after a longer training time on a more limited dataset. |
+
+## Important inference notes for the inpainting LoRAs
+
+These inpainting LoRAs were trained with a specific guide and mask setup, so input preparation during inference is important.
+
+### How to use the mask
+
+During inference, **you should not pass the mask as a separate channel**.
+
+The **mask must be embedded into the guide video**, which means:
+
+- the **mask video**
+- and the **guide video**
+
+must be treated as **a single video**.
+
+After that, you need to use the **`LTXVAddGuideMulti`** node to pass the guide video into the model.
+
+## About the mask format used during training
+
+My dataset included samples where the mask was more **blockified**. In other words, the default pattern used **8x8 blocks**.
+
+To better reproduce the training conditions during inference, you can use:
+
+- **`Blockify Mask`** from **KJNodes**
+
+This may help make the mask distribution closer to what the model saw during training.
+
+## Notes
+
+- **Base model:** `Lightricks/LTX-2.3`
+- Checkpoint behavior may vary significantly in terms of:
+  - prompt adherence
+  - use of the masked area
+  - overfitting tendency
+
+## Practical recommendations
+
+For the inpainting LoRAs in this repo:
+
+- If you want **better prompt adherence**, try the **2500 steps** checkpoint first
+- If you want **better use of the masked area**, try the **10000 steps** checkpoint first
+
+The best approach is to compare both in your workflow, since preference may vary depending on the scene, mask, and prompt.
+
+---
+
+## Examples — 2500 Steps
+
+### Example 1
+
+**Model:** `ltx23_inpaint_rank128_v1_02500steps.safetensors`  
+
+**Video:**  
+
+**Prompt:**  
+
+---
+
+### Example 2
+
+**Model:** `ltx23_inpaint_rank128_v1_02500steps.safetensors`  
+
+**Video:**  
+
+**Prompt:**  
+
+---
+
+## Examples — 10000 Steps
+
+### Example 1
+
+**Model:** `ltx23_inpaint_rank128_v1_10000steps.safetensors`  
+
+**Video:**  
+
+**Prompt:**  
+
+---
+
+### Example 2
+
+**Model:** `ltx23_inpaint_rank128_v1_10000steps.safetensors`  
+
+**Video:**  
+
+**Prompt:**  

ltx23_inpaint_rank128_v1_02500steps.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e8d2082f79be715774026a4fbbddaa2d64154d4dc9b956ade5208cc4dd8adf8
+size 1308756416

ltx23_inpaint_rank128_v1_10000steps.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f2089f5cd3cac56f93d3641c90695a7296514aa1292ec8c6f3a6ad369eda728
+size 1308756416