Trained for 91 epochs and 9600 steps.

Trained with datasets ['text-embeds-sdxl-nofilter', 'photo-concept-bucket']
Learning rate 4e-07, batch size 32, and 2 gradient accumulation steps.
Used DDPM noise scheduler for training with v_prediction prediction type and rescaled_betas_zero_snr=True
Using 'trailing' timestep spacing.
Base model: ptx0/terminus-xl-velocity-v2
VAE: madebyollin/sdxl-vae-fp16-fix

README.md

@@ -43,7 +43,7 @@ You may reuse the base model text encoder for inference.
 
 ## Training settings
 
-- Training epochs: 0
+- Training epochs: 91
 - Training steps: 9600
 - Learning rate: 4e-07
 - Effective batch size: 512
@@ -59,14 +59,6 @@ You may reuse the base model text encoder for inference.
 
 ## Datasets
 
-### dalle3
-- Repeats: 0
-- Total number of images: ~461848
-- Total number of aspect buckets: 21
-- Resolution: 1.0 megapixels
-- Cropped: False
-- Crop style: None
-- Crop aspect: None
 ### photo-concept-bucket
 - Repeats: 0
 - Total number of images: ~557568
@@ -76,3 +68,29 @@ You may reuse the base model text encoder for inference.
 - Crop style: random
 - Crop aspect: random
 
+
+## Inference
+
+
+```python
+None
+
+model_id = "terminus-xl-velocity-training"
+prompt = "a cute anime character named toast holding a sign that says SOON, sitting next to a red square on her left side, and a transparent sphere on her right side"
+negative_prompt = "malformed, disgusting, overexposed, washed-out"
+
+pipeline = DiffusionPipeline.from_pretrained(model_id)
+pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu')
+image = pipeline(
+    prompt=prompt,
+    negative_prompt='',
+    num_inference_steps=30,
+    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(1641421826),
+    width=1152,
+    height=768,
+    guidance_scale=7.5,
+    guidance_rescale=0.7,
+).images[0]
+image.save(f"output.png", format="PNG")
+```
+

model_index.json

@@ -1,19 +1,27 @@
 {
   "_class_name": "StableDiffusionXLPipeline",
-  "_diffusers_version": "0.26.0",
-  "_name_or_path": "ptx0/sdxl-base",
+  "_diffusers_version": "0.29.0.dev0",
+  "_name_or_path": "ptx0/terminus-xl-velocity-v2",
+  "feature_extractor": [
+    null,
+    null
+  ],
   "force_zeros_for_empty_prompt": true,
+  "image_encoder": [
+    null,
+    null
+  ],
   "scheduler": [
     "diffusers",
     "EulerDiscreteScheduler"
   ],
   "text_encoder": [
-    "transformers",
-    "CLIPTextModel"
+    null,
+    null
   ],
   "text_encoder_2": [
-    "transformers",
-    "CLIPTextModelWithProjection"
+    null,
+    null
   ],
   "tokenizer": [
     "transformers",

unet/config.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "UNet2DConditionModel",
   "_diffusers_version": "0.29.0.dev0",
-  "_name_or_path": "/home/user/training/lite-models/checkpoint-9500",
+  "_name_or_path": "/home/user/training/lite-models/checkpoint-9600",
   "act_fn": "silu",
   "addition_embed_type": "text_time",
   "addition_embed_type_num_heads": 64,

unet/diffusion_pytorch_model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e96ca08e5a623ca872923ade0185aa2f99fdb6c443dc8e3d9534ca9aa6f8984a
+oid sha256:843c895649bb156eadc0d9c1dadf194d991a238a556b8a645cc76b0d6214c768
 size 4994180736

unet/diffusion_pytorch_model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:517986623bd50039229e58b00b55933c4f430ea3400970a907ce034f7184302e
+oid sha256:8c134a3662ade3872068c83c11d5eade35599d091005d13f96fe89d2ab63f774
 size 140970624