Training in progress, step 20000, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1fd7698db7af5ca1d96c523abf1ce3ed39dad6f1ef364885f5698168d5072eab
+oid sha256:37a2fe2b7cf5d7145d364d49968edb54efe43efd98eba6806289d74f7e529ae8
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d356319b5cb8614b773ad954915afabc933d54cfb874f3f02d0e30cd4ce665d3
+oid sha256:f7d0dd97707c0f0f3c8a12fdba68ea5ed4110d1f028231a6ea1054c90fb8603f
 size 1544

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a96cf13c25e993c5a75039856a7a5ef57f09af02fd14f8120fb00ff3c830655
+oid sha256:59e604c620209e5ccc9d2a6df440d8eb819dc89b0c5554c52e42549c656e3250
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:aa7c19bbb017976e63ac5e9fec3eea0eda0cf5868e0275e6b00fde078e2c850d
+oid sha256:b205d8106465470d82a9f668e113b1a4c937f3fe768b385e78f85eb171e49cc2
 size 14244

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0f8f3437cd7c6266ef3d6b26d1f23a26258ce02afba475c3dabc69b494e62bdf
+oid sha256:eee3ef0a596a0ef37ef802555af22f2c3a4f5d3b4c13d00ed6df2466a73dd8ed
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:51597cfa5b705f0eeeef37eec7f4e933d1fd3032aa1f165769969345026a8d82
+oid sha256:83d15c547e182574b42460e3490faaf5cb37f299e2d2acbd532f1ef35e0b3aee
 size 1064

last-checkpoint/trainer_state.json

@@ -2,9 +2,9 @@
   "best_global_step": 18750,
   "best_metric": 1.4757392406463623,
   "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-15500",
-  "epoch": 1.4998846242596724,
+  "epoch": 1.5383432043688947,
   "eval_steps": 250,
-  "global_step": 19500,
+  "global_step": 20000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -3362,6 +3362,92 @@
       "eval_samples_per_second": 55.819,
       "eval_steps_per_second": 13.955,
       "step": 19500
+    },
+    {
+      "epoch": 1.5037304822705946,
+      "grad_norm": 0.9244908690452576,
+      "learning_rate": 0.00010110381009272005,
+      "loss": 1.3834,
+      "step": 19550
+    },
+    {
+      "epoch": 1.5075763402815168,
+      "grad_norm": 1.6488862037658691,
+      "learning_rate": 0.00010084409007090356,
+      "loss": 1.5311,
+      "step": 19600
+    },
+    {
+      "epoch": 1.511422198292439,
+      "grad_norm": 1.9130067825317383,
+      "learning_rate": 0.0001005843700490871,
+      "loss": 1.5212,
+      "step": 19650
+    },
+    {
+      "epoch": 1.5152680563033614,
+      "grad_norm": 1.326277256011963,
+      "learning_rate": 0.0001003246500272706,
+      "loss": 1.4067,
+      "step": 19700
+    },
+    {
+      "epoch": 1.5191139143142836,
+      "grad_norm": 1.7258195877075195,
+      "learning_rate": 0.00010006493000545411,
+      "loss": 1.4844,
+      "step": 19750
+    },
+    {
+      "epoch": 1.5191139143142836,
+      "eval_loss": 1.4816969633102417,
+      "eval_runtime": 17.826,
+      "eval_samples_per_second": 56.098,
+      "eval_steps_per_second": 14.024,
+      "step": 19750
+    },
+    {
+      "epoch": 1.5229597723252057,
+      "grad_norm": 1.8164838552474976,
+      "learning_rate": 9.980520998363765e-05,
+      "loss": 1.4233,
+      "step": 19800
+    },
+    {
+      "epoch": 1.5268056303361282,
+      "grad_norm": 1.5884016752243042,
+      "learning_rate": 9.954548996182115e-05,
+      "loss": 1.4313,
+      "step": 19850
+    },
+    {
+      "epoch": 1.5306514883470501,
+      "grad_norm": 1.5381648540496826,
+      "learning_rate": 9.928576994000468e-05,
+      "loss": 1.4789,
+      "step": 19900
+    },
+    {
+      "epoch": 1.5344973463579725,
+      "grad_norm": 1.6448626518249512,
+      "learning_rate": 9.90260499181882e-05,
+      "loss": 1.3756,
+      "step": 19950
+    },
+    {
+      "epoch": 1.5383432043688947,
+      "grad_norm": 1.6137230396270752,
+      "learning_rate": 9.876632989637172e-05,
+      "loss": 1.4171,
+      "step": 20000
+    },
+    {
+      "epoch": 1.5383432043688947,
+      "eval_loss": 1.4770597219467163,
+      "eval_runtime": 17.9365,
+      "eval_samples_per_second": 55.752,
+      "eval_steps_per_second": 13.938,
+      "step": 20000
     }
   ],
   "logging_steps": 50,