Trained for 1 epochs and 64000 steps.

Trained with datasets ['text-embeds', 'sfwbooru-crop']
Learning rate 1.0, batch size 6, and 1 gradient accumulation steps.
Used DDPM noise scheduler for training with epsilon prediction type and rescaled_betas_zero_snr=False
Using 'trailing' timestep spacing.
Base model: stabilityai/stable-diffusion-3.5-medium
VAE: None

README.md

@@ -329,7 +329,7 @@ You may reuse the base model text encoder for inference.
 ## Training settings
 
 - Training epochs: 1
-- Training steps: 63750
+- Training steps: 64000
 - Learning rate: 1.0
   - Learning rate schedule: cosine
   - Warmup steps: 500000

optimizer.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b23363f6adabf8db02c2c8c67006664cca9cdcfb728aa83e0bc0e0656febe09f
+oid sha256:ddb4738a7a1dc93a4c7f2dd30f5be36846175ff0c8a857f11501921af30fe470
 size 4493633377

random_states_0.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:de245d71291b56c16f75bc7bfeb91c5f3f0eaa25e1934adf1aafbad8da972a9e
-size 14408
+oid sha256:5eca79f87743cada6146b3309c797b7eedb395f7e37a8a2a1b65fe120cbd5f7b
+size 14344

scheduler.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:08edd92d79fc45bb44f24a50d7f3bfafe50bd394ffdc59aa1433e7f8a194f063
+oid sha256:d7093755a4f3ebce2e83de789544f992488222401accbe122a828334206552e7
 size 1128

training_state-sfwbooru-crop.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f9449e25be2f6ca268bdc91f4b321e29958df1364c9845bbd5f212348ecea4b9
-size 15089642
+oid sha256:57c0c31e88ebd5551d8f2a9f84963bef0aa9a3ab4c37a37480a0896141c3d570
+size 15158642

training_state.json

@@ -1 +1 @@
-{"global_step": 63750, "epoch_step": 63750, "epoch": 2, "exhausted_backends": [], "repeats": {"sfwbooru-crop": 0}}
+{"global_step": 64000, "epoch_step": 64000, "epoch": 2, "exhausted_backends": [], "repeats": {"sfwbooru-crop": 0}}

transformer/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:55f292ccdf4024692b534e7e272d1ec3fd44efc8aa4d98fd97d25f40eee4c819
+oid sha256:c7695eb5b5da1193d4460e89ae5e4aadc9cca33e6df9fcefec51ae79b0abc1f6
 size 4939433672

transformer_ema/config.json

@@ -26,7 +26,7 @@
   "min_decay": 0.0,
   "num_attention_heads": 24,
   "num_layers": 24,
-  "optimization_step": 63750,
+  "optimization_step": 64000,
   "out_channels": 16,
   "patch_size": 2,
   "pooled_projection_dim": 2048,

transformer_ema/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c826a76a6cdfcd6f77fb62d7e80f8091bc2d93f185d584859b09f8a5c003493f
+oid sha256:3ded7a841837604165d6e909a6f4f1b3f6348aef370b4fb5d8a6d0c7bdc65953
 size 9878760864

transformer_ema/ema_model.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d109bb16b001eaed043e910a2e98a1e26a2350ecefffb85f3179a6d555d6b3bc
+oid sha256:64a6ef41f961d9c3d754b0a773560dc2a1ca6c200f7fef295d93971a3eda15e3
 size 4486612394