<h1 id="blending-style">Blending - Style</h1>Blending - pictures where the background and foreground, or otherwise two separate items blend together due to having the same colour with no border to separate them.Commonly done with the character's hair being the same hue as the solid colour backdrop.A possible result of <code>no lineart</code> and <code>flat color</code> - see: <a target="_blank" rel="ugc" href="https://civarchive.com/models/1132089/flat-color-style">Flat Color - Style</a><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">ℹ️ LoRA work best when applied to the base models on which they are trained. Please read the About This Version <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">on the appropriate base models and workflow/training information.Recommended prompt structure:Positive prompt:<pre><code>blending, flat color, no lineart, negative space, {{tags}}, {{color}} background</code></pre>

Blending - Style

Trained on <a target="_blank" rel="ugc" href="https://civitai.red/models/2458426/anima?modelVersionId=2945208">Anima Base 1</a>Updated dataset, and partitioned by resolution for multi-resolution training.Using <code>captions.json</code> with 16 mixed tag/natural language variants with tag dropout and shuffle.Training config:<pre><code># trained using diffusion-pipe commit b0aa4f1e03169f3280c8518d37570a448420f8be
# NCCL_P2P_DISABLE="1" NCCL_IB_DISABLE="1" NCCL_CUMEM_ENABLE="0" deepspeed --num_gpus=1 train.py --deepspeed --config anima-lora.toml

output_dir = '/mnt/d/anima/training_output/blending-v22'

dataset = 'dataset-anima-blending.toml'

# training settings
epochs = 5
# Per-resolution batch sizes
micro_batch_size_per_gpu = [[512, 64], [1024, 32], [1280, 24]]
pipeline_stages = 1
gradient_accumulation_steps = 1
gradient_clipping = 1
warmup_steps = 100
lr_scheduler = 'cosine'

# misc settings
save_every_n_epochs = 1
activation_checkpointing = true

partition_method = 'parameters'

save_dtype = 'bfloat16'
caching_batch_size = 1
map_num_proc = 8
steps_per_print = 1
compile = true

[model]
type = 'anima'
transformer_path = '/mnt/c/workspace/models/diffusion_models/anima-base-v1.0.safetensors'
vae_path = '/mnt/c/workspace/models/vae/qwen_image_vae.safetensors'
llm_path = '/mnt/c/workspace/models/text_encoders/qwen_3_06b_base.safetensors'
dtype = 'bfloat16'
#cache_text_embeddings = false
llm_adapter_lr = 0
#timestep_sample_method = 'uniform'
flux_shift = true
multiscale_loss_weight = 0.5
sigmoid_scale = 1.3

[adapter]
type = 'lora'
rank = 32
dtype = 'bfloat16'

[optimizer]
type = 'adamw_optimi'
lr = 4e-5
betas = [0.9, 0.99]
weight_decay = 0.01
eps = 1e-8</code></pre><pre><code>resolutions = [1024, 1280, 1536]

# captions.json 16 tag+nl variants - effective 16 repeats for all images

enable_ar_bucket = true
min_ar = 0.5
max_ar = 2.0
num_ar_buckets = 9

[[directory]]
path = '/mnt/d/training_data/images_blending_update/1280x1280'
resolutions = [512, 1024, 1280]

[[directory]]
path = '/mnt/d/training_data/images_blending_update/1024x1024'
resolutions = [512, 1024]

[[directory]]
path = '/mnt/d/training_data/images_blending_update/512x512'
resolutions = [512]</code></pre>

Files with SHA256: 9da775fac9b9c968b92dfa801baee22f321fa81aa9060efad5bee11e4896b220