Instructions to use DirtScan/trail-delighter with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use DirtScan/trail-delighter with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline from diffusers.utils import load_image # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.2-klein-4B", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("DirtScan/trail-delighter") prompt = "Turn this cat into a dog" input_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png") image = pipe(image=input_image, prompt=prompt).images[0] - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
Upload trail delight Klein LoRA and training checkpoints
Browse files- training/checkpoints/trail_delight_klein4b_highres15_000000125.safetensors +3 -0
- training/checkpoints/trail_delight_klein4b_highres15_000000250.safetensors +3 -0
- training/checkpoints/trail_delight_klein4b_highres15_000000375.safetensors +3 -0
- training/config.yaml +57 -0
- training/samples/1788926920227__000000000_0.jpg +0 -0
- training/samples/1788927375069__000000125_0.jpg +0 -0
- training/samples/1788927824640__000000250_0.jpg +0 -0
- training/samples/1788928275167__000000375_0.jpg +0 -0
- training/samples/1788928721028__000000500_0.jpg +0 -0
- training/trail_delight_klein4b_highres15.safetensors +3 -0
training/checkpoints/trail_delight_klein4b_highres15_000000125.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f4fdbedc7fe72bda505c37042afe1479ccdf5178591cad46f361f3ac52af8b17
|
| 3 |
+
size 184832624
|
training/checkpoints/trail_delight_klein4b_highres15_000000250.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:605d17e09dee0fa06ce49185b2b65fefd5c371714335c24f4a2028e54d34adbd
|
| 3 |
+
size 184832624
|
training/checkpoints/trail_delight_klein4b_highres15_000000375.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2b9e43338cf75a328302b1c6687e0466a4bfaeac47186119c698e0a534b79f6
|
| 3 |
+
size 184832624
|
training/config.yaml
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
job: extension
|
| 2 |
+
config:
|
| 3 |
+
name: trail_delight_klein4b_highres15
|
| 4 |
+
process:
|
| 5 |
+
- type: diffusion_trainer
|
| 6 |
+
training_folder: /workspace/klein-lora-highres15-output
|
| 7 |
+
device: cuda
|
| 8 |
+
network:
|
| 9 |
+
type: lora
|
| 10 |
+
linear: 64
|
| 11 |
+
linear_alpha: 32
|
| 12 |
+
conv: 32
|
| 13 |
+
conv_alpha: 16
|
| 14 |
+
save:
|
| 15 |
+
dtype: bf16
|
| 16 |
+
save_every: 125
|
| 17 |
+
max_step_saves_to_keep: 4
|
| 18 |
+
datasets:
|
| 19 |
+
- folder_path: /workspace/klein-lora-highres15-dataset/target
|
| 20 |
+
control_path: /workspace/klein-lora-highres15-dataset/reference
|
| 21 |
+
caption_ext: txt
|
| 22 |
+
resolution:
|
| 23 |
+
- 512
|
| 24 |
+
train:
|
| 25 |
+
batch_size: 1
|
| 26 |
+
steps: 500
|
| 27 |
+
gradient_accumulation_steps: 1
|
| 28 |
+
train_unet: true
|
| 29 |
+
train_text_encoder: false
|
| 30 |
+
gradient_checkpointing: true
|
| 31 |
+
noise_scheduler: flowmatch
|
| 32 |
+
optimizer: adamw8bit
|
| 33 |
+
lr: 0.0001
|
| 34 |
+
timestep_type: weighted
|
| 35 |
+
content_or_style: balanced
|
| 36 |
+
dtype: bf16
|
| 37 |
+
model:
|
| 38 |
+
name_or_path: black-forest-labs/FLUX.2-klein-base-4B
|
| 39 |
+
arch: flux2_klein_4b
|
| 40 |
+
quantize: true
|
| 41 |
+
low_vram: false
|
| 42 |
+
sample:
|
| 43 |
+
sampler: flowmatch
|
| 44 |
+
sample_every: 125
|
| 45 |
+
sample_start_step: 0
|
| 46 |
+
width: 768
|
| 47 |
+
height: 448
|
| 48 |
+
prompts:
|
| 49 |
+
- Remove cast shadows and glare from this trail-surface photograph. Preserve
|
| 50 |
+
the exact trail geometry, rocks, leaves, markings, and camera viewpoint. Use
|
| 51 |
+
neutral soft overcast illumination.
|
| 52 |
+
seed: 0
|
| 53 |
+
guidance_scale: 1
|
| 54 |
+
sample_steps: 4
|
| 55 |
+
meta:
|
| 56 |
+
name: trail_delight_klein4b_highres15
|
| 57 |
+
version: '1.0'
|
training/samples/1788926920227__000000000_0.jpg
ADDED
|
training/samples/1788927375069__000000125_0.jpg
ADDED
|
training/samples/1788927824640__000000250_0.jpg
ADDED
|
training/samples/1788928275167__000000375_0.jpg
ADDED
|
training/samples/1788928721028__000000500_0.jpg
ADDED
|
training/trail_delight_klein4b_highres15.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e7b145953b056865bdb9f664005ddeee27f7652753e7586904b7f8befe14ff60
|
| 3 |
+
size 184832624
|