Training in progress, step 38000, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:23dcc181cec0e10d847e208901d6626875ec6b62c6c0f26d433995f3783e85d0
+oid sha256:6a045cd1651964484e35a7ee4cc80f394400bcf32d399ca366f907f0dafb8b29
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7432e9c2d93150abfd716acce42331362afc030b8c095403869cbf1291225746
+oid sha256:47093e48c8dcb6ad57ae3eda9ce7141b8c756fad2f39a70827cbbb9ed49f5649
 size 1544

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1b8434a82bb738c0cecb4adf19f6cc16b50060235313456c478572cc44126b33
+oid sha256:b60bdca21a24c4c9f76b11b903c37c4127fc1f54a299af327f2fe260c58ff720
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c5a317d14f0c4fea42d2b4ca5f020a67142d647545709c1c9006f254a7d7349
-size 14308
+oid sha256:0df6fc486b7517363afca4fd7698726ecbf63a16d3ce84e06ce60f5376cd9b7e
+size 14244

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4d13c8e281c1623531d4d497478b975ea082917bb2c61878d98cb5fb27c6af9
+oid sha256:97243af249b903bf91950a8fcb37e9bd58551a34f032c66a6b68e32bc4197313
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:23ce717bb1ff7d19ce8b39673c5e006d14b3fec124190d834c88a63ab05da6d0
+oid sha256:541cb6dea9c80e2359e5134078f161ec83a5de3af8e29b3e9b6b880e5dd9058f
 size 1064

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 36750,
-  "best_metric": 0.9847651720046997,
-  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-34000",
-  "epoch": 2.8843935081916774,
+  "best_global_step": 38000,
+  "best_metric": 0.98431396484375,
+  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-38000",
+  "epoch": 2.9228520883008997,
   "eval_steps": 250,
-  "global_step": 37500,
+  "global_step": 38000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -6458,6 +6458,92 @@
       "eval_samples_per_second": 58.059,
       "eval_steps_per_second": 14.515,
       "step": 37500
+    },
+    {
+      "epoch": 2.8882393662026,
+      "grad_norm": 0.6800574660301208,
+      "learning_rate": 7.630574240968237e-06,
+      "loss": 0.983,
+      "step": 37550
+    },
+    {
+      "epoch": 2.892085224213522,
+      "grad_norm": 0.9417561292648315,
+      "learning_rate": 7.3708542191517545e-06,
+      "loss": 0.973,
+      "step": 37600
+    },
+    {
+      "epoch": 2.895931082224444,
+      "grad_norm": 0.6454310417175293,
+      "learning_rate": 7.111134197335273e-06,
+      "loss": 1.0493,
+      "step": 37650
+    },
+    {
+      "epoch": 2.8997769402353666,
+      "grad_norm": 1.0287562608718872,
+      "learning_rate": 6.851414175518792e-06,
+      "loss": 0.9821,
+      "step": 37700
+    },
+    {
+      "epoch": 2.903622798246289,
+      "grad_norm": 1.6874816417694092,
+      "learning_rate": 6.591694153702309e-06,
+      "loss": 0.9761,
+      "step": 37750
+    },
+    {
+      "epoch": 2.903622798246289,
+      "eval_loss": 0.9886119961738586,
+      "eval_runtime": 17.2542,
+      "eval_samples_per_second": 57.957,
+      "eval_steps_per_second": 14.489,
+      "step": 37750
+    },
+    {
+      "epoch": 2.907468656257211,
+      "grad_norm": 0.7677489519119263,
+      "learning_rate": 6.331974131885828e-06,
+      "loss": 1.0421,
+      "step": 37800
+    },
+    {
+      "epoch": 2.9113145142681334,
+      "grad_norm": 0.9784395098686218,
+      "learning_rate": 6.072254110069346e-06,
+      "loss": 0.9288,
+      "step": 37850
+    },
+    {
+      "epoch": 2.9151603722790553,
+      "grad_norm": 0.6236763000488281,
+      "learning_rate": 5.812534088252864e-06,
+      "loss": 0.9777,
+      "step": 37900
+    },
+    {
+      "epoch": 2.9190062302899777,
+      "grad_norm": 1.0675079822540283,
+      "learning_rate": 5.552814066436382e-06,
+      "loss": 0.9659,
+      "step": 37950
+    },
+    {
+      "epoch": 2.9228520883008997,
+      "grad_norm": 0.9310262203216553,
+      "learning_rate": 5.2930940446198996e-06,
+      "loss": 0.9334,
+      "step": 38000
+    },
+    {
+      "epoch": 2.9228520883008997,
+      "eval_loss": 0.98431396484375,
+      "eval_runtime": 17.3967,
+      "eval_samples_per_second": 57.482,
+      "eval_steps_per_second": 14.371,
+      "step": 38000
     }
   ],
   "logging_steps": 50,