Training in progress, step 22000, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:abe147d1fd4bc110db7089e008a574b65f554aaa98420567a0bc2b63656bd9c1
+oid sha256:300bf30417bbf3f32527e2186c9abed7ed76183b832bf2d5e14eac86f900b976
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a3abcac0bf194b5337811637955f6e6b362380aa1f532202481201a600657324
+oid sha256:63292d3528a5a0926c722da43abe6585432ca093708fd76551d8031e7ddd3643
 size 1736

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d206c147f3ebf5a17a6c44dea5417a1ae1dd50a6fd22e30c6aea36175324b86
+oid sha256:bd93d8bd898fd17bb0bb48e446465b209d65cfac9f90818590fb648e18b8ab7a
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4a5d062bb7d441f36a44929472f10d0b795d7b5f5a9dcd38edd739ffa77a6d36
-size 14180
+oid sha256:ec8bd4a1e6303e385dc1e7236df83c38e8a5367849e5bc57e5960781558c1b9e
+size 14244

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e00a972afc33ed5290e59d706523471e5c191feb09759aef8aeb13be22ede6e8
+oid sha256:f1c27ce326198bdf77c595dcbaf5251be4cae7f085aadb1446913e62d977b694
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:afb6a0769e42f6963785d597ff6271423cd7dd782771f3b3c0a05c348f1cd005
+oid sha256:aef577ff6c2918592ffb94dbf847593611b785b48ba36e1fc639ca9ece0452f2
 size 1064

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 21500,
-  "best_metric": 0.77164626121521,
+  "best_global_step": 21750,
+  "best_metric": 0.7681905031204224,
   "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-21500",
-  "epoch": 1.6537189446965619,
+  "epoch": 1.6921775248057842,
   "eval_steps": 250,
-  "global_step": 21500,
+  "global_step": 22000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -3706,6 +3706,92 @@
       "eval_samples_per_second": 55.765,
       "eval_steps_per_second": 13.941,
       "step": 21500
+    },
+    {
+      "epoch": 1.657564802707484,
+      "grad_norm": 0.8676180243492126,
+      "learning_rate": 9.070462041918813e-05,
+      "loss": 0.8113,
+      "step": 21550
+    },
+    {
+      "epoch": 1.6614106607184063,
+      "grad_norm": 0.6232962012290955,
+      "learning_rate": 9.044490039737163e-05,
+      "loss": 0.7847,
+      "step": 21600
+    },
+    {
+      "epoch": 1.6652565187293287,
+      "grad_norm": 1.151957631111145,
+      "learning_rate": 9.018518037555515e-05,
+      "loss": 0.7924,
+      "step": 21650
+    },
+    {
+      "epoch": 1.6691023767402506,
+      "grad_norm": 0.8149247169494629,
+      "learning_rate": 8.992546035373868e-05,
+      "loss": 0.7395,
+      "step": 21700
+    },
+    {
+      "epoch": 1.672948234751173,
+      "grad_norm": 1.0983916521072388,
+      "learning_rate": 8.96657403319222e-05,
+      "loss": 0.7715,
+      "step": 21750
+    },
+    {
+      "epoch": 1.672948234751173,
+      "eval_loss": 0.7681905031204224,
+      "eval_runtime": 17.9279,
+      "eval_samples_per_second": 55.779,
+      "eval_steps_per_second": 13.945,
+      "step": 21750
+    },
+    {
+      "epoch": 1.6767940927620952,
+      "grad_norm": 0.7059574723243713,
+      "learning_rate": 8.94060203101057e-05,
+      "loss": 0.784,
+      "step": 21800
+    },
+    {
+      "epoch": 1.6806399507730174,
+      "grad_norm": 1.0587022304534912,
+      "learning_rate": 8.914630028828923e-05,
+      "loss": 0.7608,
+      "step": 21850
+    },
+    {
+      "epoch": 1.6844858087839398,
+      "grad_norm": 0.7638582587242126,
+      "learning_rate": 8.888658026647275e-05,
+      "loss": 0.8159,
+      "step": 21900
+    },
+    {
+      "epoch": 1.6883316667948618,
+      "grad_norm": 0.5783549547195435,
+      "learning_rate": 8.862686024465626e-05,
+      "loss": 0.7959,
+      "step": 21950
+    },
+    {
+      "epoch": 1.6921775248057842,
+      "grad_norm": 1.2192896604537964,
+      "learning_rate": 8.836714022283978e-05,
+      "loss": 0.8272,
+      "step": 22000
+    },
+    {
+      "epoch": 1.6921775248057842,
+      "eval_loss": 0.7701475024223328,
+      "eval_runtime": 17.8145,
+      "eval_samples_per_second": 56.134,
+      "eval_steps_per_second": 14.034,
+      "step": 22000
     }
   ],
   "logging_steps": 50,