README.md

---
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- text: >-
    A photo of a man in a brightly colored tracksuit and gold chains, captured
    mid-power move while breakdancing on wet pavement under neon lights, y2k
    style photo with reflections.
  output:
    url: images/ComfyUI_00002_.png
- text: >-
    act1vate a photo of a woman in chunky platform boots dynamically mid-splash
    landing in a puddle (foreground splash), rainy city street reflecting neon
    in the midground, blurred car headlights in the background, y2k style.
  output:
    url: images/ComfyUI_00035_.png
- text: >-
    A photo of a man wearing a basketball jersey over a hoodie, caught mid-air
    executing a complex dunk on an urban court, dynamic fisheye shot from below
    the hoop.
  output:
    url: images/ComfyUI_00006_.png
- text: >-
    A photo of a woman wearing velour tracksuit bottoms and sunglasses,
    dynamically leaning out of a speeding convertible, hair whipping in the
    wind, low-angle action shot.
  output:
    url: images/ComfyUI_00007_.png
- text: >-
    A photo of a man in extremely baggy jeans, dynamically sliding down a long
    handrail on rollerblades, low-angle y2k action photograph.
  output:
    url: images/ComfyUI_00012_.png
- text: >-
    A photo of a man wearing a fuzzy bucket hat, dynamically posed riding a BMX
    bike on its back wheel (wheelie) down a city street, low-angle shot.
  output:
    url: images/ComfyUI_00014_.png
- text: >-
    act1vate a photo of a woman in baggy cargo pants dynamically mid-kickflip on
    a skateboard in the foreground, with steps receding into the midground, and
    blurred city traffic visible in the background distance, captured via
    low-angle fisheye.
  output:
    url: images/ComfyUI_00021_.png
- text: >-
    act1vate a photo of a man in a vibrant tracksuit breakdancing powerfully in
    the midground on wet pavement reflecting neon signs from the background,
    foreground showing discarded flyers, dynamic y2k style fisheye shot.
  output:
    url: images/ComfyUI_00022_.png
- text: >-
    act1vate a photo of a woman with windswept hair dynamically leaning far out
    of a convertible (midground), the car door detailed in the foreground, a
    coastal highway stretching into the background sunset, low-angle action
    shot.
  output:
    url: images/ComfyUI_00027_.png
- text: >-
    act1vate a photo of a man striking a dynamic pose with a large boombox on
    his shoulder in the foreground, standing before a midground wall layered
    with peeling posters, an alleyway receding into darkness in the background,
    y2k style fisheye.
  output:
    url: images/ComfyUI_00028_.png
- text: >-
    act1vate a photo of a man dynamically mid-vault over a subway turnstile
    (foreground barrier), the tiled station platform stretching into the
    midground, train lights approaching from the background tunnel, low-angle
    capture.
  output:
    url: images/ComfyUI_00030_.png
- text: >-
    act1vate a photo of a man holding an old digital camera pointed outwards
    (foreground object), dynamically posing against a wall of vintage TVs
    displaying static (midground), shop interior blurred in the background,
    low-angle fisheye.
  output:
    url: images/ComfyUI_00036_.png
base_model: black-forest-labs/FLUX.1-dev
instance_prompt: act1vate, fisheye, dynamic pose
license: mit
---
# ActionLens

<Gallery />

## Model description 

A lora I trained off of various images with interesting poses and with a slight fisheye lens. mainly the training data included y2k photography shots. It&#39;s incredibly flawed often messing up hands and details but creates really good starting points to work with in krita and iterate off of. I didn&#39;t want to lose it through updates so uploading it.

recommended strength: 0.8 when using alone - 1 when combined with other loras

Activation words and phrases: act1vate, fisheye, dynamic pose

## Trigger words

You should use `act1vate` to trigger the image generation.

You should use `fisheye` to trigger the image generation.

You should use `dynamic pose` to trigger the image generation.


## Download model

Weights for this model are available in Safetensors format.

[Download](/CtrlAltArt/Flux_ActionLens/tree/main) them in the Files & versions tab.