Trained for 2 epochs and 3000 steps.

Trained with datasets ['text-embeds', 'mj-v6']
Learning rate 8e-06, batch size 2, 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: ptx0/pixart-900m-1024-ft
VAE: madebyollin/sdxl-vae-fp16-fix

README.md

@@ -211,7 +211,7 @@ You may reuse the base model text encoder for inference.
 ## Training settings
 
 - Training epochs: 2
-- Training steps: 2500
+- Training steps: 3000
 - Learning rate: 8e-06
 - Effective batch size: 2
   - Micro-batch size: 2

optimizer.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0f8fea7662313eec7cb12f967f20148323ad21b65624783d26792e6423c44673
+oid sha256:ab3d25f3b9f41c0b12e3f59e609f7fe6b2b16bf1009f49a1375010f3f702c485
 size 5451415117

random_states_0.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d8f06e19ca759e2a76cf92bbbb08934473c201976326df04622c4fc80923bf6b
-size 14344
+oid sha256:f878f772eceda6e47376fb7f6ed4b60eadac7bf94ad1b1b00d71102578e5eddc
+size 14408

scheduler.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d394596b6a9a6eb220c4b514546efe6b5d39875b7f3c514eaa13e2af3c14e8d3
+oid sha256:4beec63126998040f00cefceadff78db11f493a3f2cc333822a34fdde9434243
 size 1000

training_state.json

@@ -1 +1 @@
-{"global_step": 2500, "epoch_step": 320, "epoch": 3, "exhausted_backends": [], "repeats": {"mj-v6": 0}}
+{"global_step": 3000, "epoch_step": 820, "epoch": 3, "exhausted_backends": [], "repeats": {"mj-v6": 0}}

transformer/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:411a2107d308c0db404e09f867ab837fa68aac9ed6d2708dbe9e42d421d91d4b
+oid sha256:fb807865ba5cfced0d63e19c430c911ca7038714c534cfbf4a4a6e4bc7344603
 size 1816969728