Trained for 3 epochs and 4000 steps.

Trained with datasets ['text-embeds-pixart-nofilter', 'photo-concept-bucket']
Learning rate 4e-07, batch size 32, and 2 gradient accumulation steps.
Used DDPM noise scheduler for training with epsilon prediction type and rescaled_betas_zero_snr=False
Using 'linspace' timestep spacing.
Base model: terminusresearch/pixart-900m-1024
VAE: madebyollin/sdxl-vae-fp16-fix

README.md

@@ -1006,7 +1006,7 @@ You may reuse the base model text encoder for inference.
 ## Training settings
 
 - Training epochs: 3
-- Training steps: 3500
+- Training steps: 4000
 - Learning rate: 4e-07
 - Effective batch size: 512
   - Micro-batch size: 32

optimizer.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4edaaf19920d3d6a46607c022d3b6b470ec450aaf717922999e16391b0b1b46e
+oid sha256:d3f2a207e9d14a0fdf4a0a7145ebb9755a5cafe679787f1614dfe79ff3413514
 size 5451415117

random_states_0.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1495b449207bdadd4ea0402389981ad0a4d56b4e73903a636cc36f0d41548c26
+oid sha256:e7bd4628d8a964d3f04e0fcb113e90248752b7e241a765b3bdb1f93f14da0fe9
 size 16100

scheduler.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:636a5983c58cb34c8629f71677861a6d1f13f0369fb3fbe0365526df69f97f6a
+oid sha256:d54f64f68660d822e560095ef18c71568b1d07fc2e99f70252173d5b305e6cb6
 size 1000

training_state-photo-concept-bucket.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a3a37ef39f79119c29bd9150bbe4d8becca363bb2e476841b4a5f68fda61faf
-size 5364416
+oid sha256:df90f1ee82d73418b70af1126a8f49e0a53fbb8c20218f5fd65d2d1bddf56f2a
+size 7561740

training_state.json

@@ -1 +1 @@
-{"global_step": 3500, "epoch_step": 224, "epoch": 4, "exhausted_backends": [], "repeats": {"photo-concept-bucket": 0}}
+{"global_step": 4000, "epoch_step": 724, "epoch": 4, "exhausted_backends": [], "repeats": {"photo-concept-bucket": 0}}

transformer/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:207d463be1d711e9f7466b69c1bf137ea0a09b860c6af833155dee9e09e9bf6b
+oid sha256:dd7bcb5bc3bb39c5a9afd9bb2e50dfccdb07e00d95f4ea2066f9b9db52446908
 size 1816969728