LCM Hyperloop - 8 steps - 14 sec - 12 images — hyperloop-v37.8
images of 1536 x 2688 in 63 seconds!
Runtime profile
- Access
- External workflow
- Status
- Listed
- Version
- HyperLoop-v37.8.json
- Source
- Civitai
- Updated
- Sep 30, 2026
Source description
images of 1536 x 2688 in 63 seconds!
updates: much better output quality
We added a model merge, which improved the output quality, and now model options are not restricted to Cardos and Nightvision; other models are getting good results as well.
the work flow is now is 10 steps instead of 8, i have seen differency between my GPUs, one does better with 8 steps "4-4" the other does better with 10 "5-5"
2 IP adapters are now IPadapter Plus, and "IPAdapter Plus Face" is separate, having its own node
Added 1 additional "Apply node" for for each IP Adapeter to have more control over the Base model and refiner model IPadapter weight
I forgot to name the "apply nodes for IP Adapters Base and Refiner, The base model control on top, the refiner model control is below.
To improve the speed, change the save image nodes in the work flow to the standard save image node; it will improve at least 16+ seconds for 6 images.
This version's Batch size reduced to 6 images, with an image size of 768x1344, then upscaling to 1536x2688
SDXL Upscaling using RTX4090
(1536x1536 upscale size, not image size) took 63 seconds for 6 images, double the SDXL recommended size
(1024x1024 upscale size, not image size) took 81 secondsfor 6 images, double the SDXL recommended size
12 images in 2 minutes and 1.4 seconds! 1536x2688....
Canva size: 1536x2688 Image size: 768x1344 Scale factor: 2 Upscaling iteration 1 with scale factor 2 Tile size: 1536x1536 Tiles amount: 2 Grid: 2x1 Redraw enabled: True Seams fix mode: NONE Requested to load SDXL Loading 1 new model 100%|████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:16
SD1.5 Upscaling:
I didn't do the speed test as i am still working to improve it
What you need
- Model files
- 5.4 GiB known subtotal
Remaining hardware or download-size estimates are not available yet.
Requirements
Models
19- Found
Upscaler · 63.9 MB · Hugging Face · Kim2091/UltraSharp
add-detail-xl.safetensors
Not resolvedCheckpoint · Unknown
AdvancedEnhancerXLv2.safetensors
Not resolvedCheckpoint · Unknown
albedobaseXL_v10.safetensors
Not resolvedCheckpoint · Unknown
Checkpoint · 2.22 GB · SAFETENSORS · fp16 · pruned · Civitai · 8030
Andromeda_5-000002.safetensors
Not resolvedCheckpoint · Unknown
- Found
Object Detector · 5.94 MB · Hugging Face · Bingsu/adetailer
binarycode.safetensors
Not resolvedCheckpoint · Unknown
control_v11f1e_sd15_tile.pth
Not resolvedControlNet · Unknown
LoRA · 967 MB · Hugging Face · mr-Dan/ip-adapter-plus_sdxl_vit-h
LoRA · 967 MB · Hugging Face · mr-Dan/ip-adapter-plus-face_sdxl_vit-h
lcm_1.5_lora_weights.safetensors
Not resolvedLoRA · Unknown
lcm_sdxl_lora.safetensors
Not resolvedLoRA · Unknown
- Found
Vision model · 850 MB · Hugging Face · google-t5/t5-base
nightvisionXLPhotorealisticPortrait_release0770Bakedvae.safetensors
Not resolvedVAE · Unknown
RPGFireElementalXL.safetensors
Not resolvedCheckpoint · Unknown
- Found
Segmentation · 358 MB · PTH · Hugging Face · comfyuistudio/sam_vit_b_01ec64.pth
SD1.5-SRv0.1.safetensors
Not resolvedCheckpoint · Unknown
SR-loconT1-hAI-XL-000005.safetensors
Not resolvedCheckpoint · Unknown
AI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
Purpose
LCM Hyperloop is an SDXL 1.0 image-generation workflow with text and loaded-image inputs, saved image outputs, and a described option to upscale 768 × 1344 images 2× to 1536 × 2688.
Flow
The described settings batch six 768 × 1344 images and upscale them 2× in one iteration to 1536 × 2688, using two 1536 × 1536 tiles in a 2 × 1 grid; redraw is enabled and seam fixing is off.
Inputs and outputs
The listed workflow includes Text and LoadImage nodes, plus an Image Save node.
Setup
For a lower-powered GPU, the author advises reducing the batch size; the example describes using an RTX 4090.
Dependencies
The workflow lists 4x-UltraSharp.pth, add-detail-xl.safetensors, AdvancedEnhancerXLv2.safetensors, albedobaseXL_v10.safetensors, analogMadness_v50.safetensors, Andromeda_5-000002.safetensors, and bbox/face_yolov8n.pt.
Other listed files are binarycode.safetensors, control_v11f1e_sd15_tile.pth, ip-adapter-plus_sdxl_vit-h.bin, ip-adapter-plus-face_sdxl_vit-h.bin, lcm_1.5_lora_weights.safetensors, and lcm_sdxl_lora.safetensors.
The remaining listed files are model.safetensors, nightvisionXLPhotorealisticPortrait_release0770Bakedvae.safetensors, RPGFireElementalXL.safetensors, sam_vit_b_01ec64.pth, SD1.5-SRv0.1.safetensors, and SR-loconT1-hAI-XL-000005.safetensors.
Configuration
The author’s example uses Cardos XL v1 as the base model and NightVision XL - Photorealistic 0.7.7.0 as the refiner; later notes say other models can also be used.
IPAdapter Plus and IPAdapter Plus Face have separate nodes and Apply controls for base-model and refiner weights; the base control is above the refiner control.
The author reports different step preferences on different GPUs: 8 steps split 4–4 on one, and 10 steps split 5–5 on another.
Cautions
Check the loaded settings: the title says 8 steps and 12 images, while the update describes 10 steps and a batch of six.
Suggestion
No license is listed; check 7sanal7san’s usage terms before use.
Suggestion
Does this need to be edited?
Sign in to send an edit request.
Sources
1 source
Sources
1 sourceSource excerpts
3 excerpts
Source excerpts
3 excerptsSource context: 1257 downloads · Type Workflows · Base model SDXL 1.0
Source context: 1255 downloads · Type Workflows · Base model SDXL 1.0
Source context: 1191 downloads · Type Workflows · Base model SDXL 1.0
