Hi folks, this is CCS, and today we’ve got a double-upgrade drop — two huge updates to the tools powering your Qwen and WAN workflows:
➜ IAMCCS_QE_PROMPT_ENHANCER v1.0.1
👉 https://github.com/IAMCCS/IAMCCS_QE_PROMPT_ENHANCER
➜ IAMCCS_nodes v1.3.0 (major update)
👉 https://github.com/IAMCCS/IAMCCS_nodes
This is the kind of update that doesn’t just “fix things”…
It makes your entire pipeline smarter, quicker, and more VRAM-friendly.
Let’s go.
⭐ 1 — IAMCCS_QE_PROMPT_ENHANCER v1.0.1
What’s new in 1.0.1?
🆕 UI Upgrade (polished, faster, clearer)
-
New toggle bar for LoRA modes & QwenVL modes
-
Better highlight behavior and synced dropdown/preset selection
-
Cleaner layout → easier preset browsing
-
New dark thumbnails for better readability
🎥 New Camera Angle Presets
The Camera Angle category finally got the cinematic treatment:
-
2 wide angles preset
-
2 close up preset
-
new chinese translations
These presets generate Qwen-friendly semantic instructions, not generic blurbs.
Perfect for storyboarding, cinematics, and consistent editing styles.
check my previous post to see this node in action!
https://www.patreon.com/posts/lets-try-qwen-142884126?utm_medium=clipboard_copy&utm_source=copyLink&utm_campaign=postshare_creator&utm_content=join_link
⚠️ Low VRAM Users — Important Note
NUNCHAKU USERS: If your GPU sits below 12 GB VRAM:
➡️ Use the updated Qwen LoRA Loader inside IAMCCS_nodes (v1.3.0).
It ensures Qwen LoRAs load without VRAM overflow and works flawlessly with Nunchaku models.
You can grab IAMCCS_nodes here:
👉 https://github.com/IAMCCS/IAMCCS_nodes
⭐ 2 — IAMCCS_nodes v1.3.0 (Massive Upgrade)
Version 1.3.0 is big.
This update expands the toolset, cleans the internal logic, and introduces completely new nodes built for both WANAnimate/WAN2.2 and Qwen Image Edit pipelines.
🆕 1 — LoRA Stack (MODEL In→Out) WAN
A brand-new node that applies up to 4 WAN/Flow/Standard LoRAs directly onto a loaded MODEL — in a single step (model in-out pipeline).
Perfect for:
Why this matters:
Keeps your graph clean.
Reduces extra hops.
Simplifies LoRA lineage in complex videos.
🆕 2 — Qwen Nunchaku Image LoRA Loader FIX (for Low VRAM)
This is the node everyone using Qwen Nunchaku needed — especially under 12GB VRAM.
Highlights:
-
Fixes LoRA loading for Qwen Image Edit (Nunchaku)
-
Auto-wraps transformers and safely merges LoRA deltas
-
Full VRAM-aware offload system
-
Perfect for Nunchaku Qwen setups
-
Finally allows Qwen Nunchaku LoRA use on low VRAM GPUs
This node alone can save you hundreds of VRAM errors.
🆕 3 — Internal WAN Key Remapping (2.1 → 2.2 cross support)
LoRA weight shapes change between WAN versions.
This update now handles everything automatically, including:
-
LightX2V
-
WAN 2.1 / 2.2
-
4-step LoRAs
-
Character LoRAs
-
Multi-LoRA stacks
No more broken keys.
No more half-loaded LoRA weights.
⭐ 3 — IAMCCS_annotate update
Some words about my little baby…
Just a small update about my little baby…
It’s now available both on the repo and through the ComfyUI Manager search. 🚀
👉 https://github.com/IAMCCS/IAMCCS_annotate
⭐ Final Words
Both updates (1.0.1 + 1.3.0) were built after weeks of testing inside real workflows — face swap, cinematic LUT edits, pose-driven changes, relighting, and multi-image pipelines.
If you’re editing with Qwen or animating with WAN, these two updates will literally make your life easier.
Stay tuned 4 more ❤️