Trained for 5 epochs and 5500 steps.

Trained with datasets ['text-embeds', 'mj-v6']
Learning rate 8e-06, batch size 32, and 4 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: PixArt-alpha/PixArt-Sigma-XL-2-1024-MS
VAE: madebyollin/sdxl-vae-fp16-fix

README.md

@@ -210,8 +210,8 @@ You may reuse the base model text encoder for inference.
 
 ## Training settings
 
-- Training epochs: 4
-- Training steps: 5000
+- Training epochs: 5
+- Training steps: 5500
 - Learning rate: 8e-06
 - Effective batch size: 128
   - Micro-batch size: 32

optimizer.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b408c484f29623d82441f52f58f2d45b9c225c2d747b330624aad2474df1520
+oid sha256:083ced81d053298275b56202c9e468c054311eaf5c0de4ecd8c2fbe4948b09a9
 size 3665677155

random_states_0.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:86ae4043f57450c7708f7daf759c419c4397cd95df061dc1b051de557a991eec
-size 14344
+oid sha256:f0fddde73d47b8aafd98e3c6c125faf1eea661799c7eacbb92771d93d8ee006a
+size 14280

scheduler.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:08d2b1d2ff87fc142d6b57325846dcb28ce085536f556fb63ed5e84f3e0ddc37
+oid sha256:cbec9aff97e801a07f80ddafe16983f4f0ad6cbed9817c950ad8294236d1685d
 size 1000

training_state-mj-v6.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3cef1b05385597e3baa55f8913fca926efa04a3b3c3d356f85f3ee49877635e8
-size 16049251
+oid sha256:60b5b681d9bef75b790f5743c5fe332600a5777e22bfefaaa77e2611a2e364cd
+size 10297443

training_state.json

@@ -1 +1 @@
-{"global_step": 5000, "epoch_step": 557, "epoch": 5, "exhausted_backends": [], "repeats": {"mj-v6": 0}}
+{"global_step": 5500, "epoch_step": 9, "epoch": 6, "exhausted_backends": [], "repeats": {"mj-v6": 0}}

transformer/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:046425dd53749c2893bc16e71223ece74b59d2e1fcfe475c5063c31753c33fd5
+oid sha256:388dd63045347e4c7f9d1a967b8278c958ea468811be170f5552d3bdd5b94bdb
 size 1221780352