Upload 2695779/3032778/README.md with huggingface_hub
Browse files- 2695779/3032778/README.md +20 -9
2695779/3032778/README.md
CHANGED
|
@@ -36,32 +36,43 @@ tags:
|
|
| 36 |
|
| 37 |
A double handjob blowbang.
|
| 38 |
|
| 39 |
-
Include the keyword **bl0wb4ng** in your prompt and wan will handle the rest.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
|
| 41 |
Usage
|
| 42 |
-----
|
| 43 |
|
| 44 |
-
* **Prompting**: keep it simple.
|
| 45 |
* **Shift**: keep it low. **7** is OK with most of schedulers and evenly split steps.
|
| 46 |
* **Cut**: the cut happens at the 9th frame.
|
| 47 |
* Lower **Strengths** (high-noise):
|
| 48 |
|
| 49 |
+ ~0.5: makes a progressive transition instead of a cut.
|
| 50 |
-
+ ~0.7:
|
| 51 |
|
| 52 |
Misc
|
| 53 |
----
|
| 54 |
|
| 55 |
-
All the showcase videos have their **workflow** embedded.
|
| 56 |
-
|
| 57 |
-
Tested only with **lightx2v** LoRAs enabled (1030 on high-noise and 1022 on low-noise).
|
| 58 |
|
| 59 |
-
|
|
|
|
|
|
|
| 60 |
|
| 61 |
Training
|
| 62 |
--------
|
| 63 |
|
| 64 |
-
67.5 epochs, 72 5-seconds clips (4860 steps).
|
|
|
|
| 65 |
|
| 66 |
---
|
| 67 |
|
|
@@ -71,4 +82,4 @@ Model: [CivitAI Model Page](https://civitai.com/models/2695779?modelVersionId=30
|
|
| 71 |
|
| 72 |
Archive: [CivArchive Page](https://civarchive.com/models/2695779?modelVersionId=3032778)
|
| 73 |
|
| 74 |
-
<!-- Version:
|
|
|
|
| 36 |
|
| 37 |
A double handjob blowbang.
|
| 38 |
|
| 39 |
+
Include the keyword **bl0wb4ng** in your prompt and wan will *handle* the rest.
|
| 40 |
+
|
| 41 |
+
v1.1
|
| 42 |
+
====
|
| 43 |
+
|
| 44 |
+
Tweaked a few knobs and pushed the training a bit further.
|
| 45 |
+
|
| 46 |
+
🟢 Less hallucinations
|
| 47 |
+
🟢 Better hand skills
|
| 48 |
+
🔴 Slightly reduced facial accuracy
|
| 49 |
|
| 50 |
Usage
|
| 51 |
-----
|
| 52 |
|
| 53 |
+
* **Prompting**: keep it simple; avoid NSFW vocabulary when possible.
|
| 54 |
* **Shift**: keep it low. **7** is OK with most of schedulers and evenly split steps.
|
| 55 |
* **Cut**: the cut happens at the 9th frame.
|
| 56 |
* Lower **Strengths** (high-noise):
|
| 57 |
|
| 58 |
+ ~0.5: makes a progressive transition instead of a cut.
|
| 59 |
+
+ ~0.7: less crowded scene.
|
| 60 |
|
| 61 |
Misc
|
| 62 |
----
|
| 63 |
|
| 64 |
+
All the showcase videos have their **workflow** embedded.
|
| 65 |
+
Tested only:
|
|
|
|
| 66 |
|
| 67 |
+
* with **lightx2v** LoRAs enabled (1030 or 1022 on high-noise and 1022 on low-noise).
|
| 68 |
+
* in **ComfyUI**.
|
| 69 |
+
* on the following **base wan** models: Q4, Q5, Q6 and nvfp4.
|
| 70 |
|
| 71 |
Training
|
| 72 |
--------
|
| 73 |
|
| 74 |
+
* v1.0: 67.5 epochs, 72 5-seconds clips (4860 steps).
|
| 75 |
+
* v1.1: 80.0 epochs, 72 5-seconds clips (5760 steps).
|
| 76 |
|
| 77 |
---
|
| 78 |
|
|
|
|
| 82 |
|
| 83 |
Archive: [CivArchive Page](https://civarchive.com/models/2695779?modelVersionId=3032778)
|
| 84 |
|
| 85 |
+
<!-- Version: 20260617_upload -->
|