Training in progress, step 21000, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dfd510b5f0a4135033a16f697b55c6f68a44bbbbde64f2deabad3983170c54fe
+oid sha256:3eb66659915116c9f52f8b7d65982f16dfe01a2f397d774e7e2e124b2fee93e6
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:476394277ce8fd3dba6a07af0c9c863330e18dfb02652447fd61d7d7f00046ff
+oid sha256:094124f70673a425259be33a9a30631e505be7e6b457e57a5337a37485da2199
 size 1736

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b5aba9078e1bdd7ec7d7b7ae2aa74c57192f6519a73584e6ce2a47c0539cdcbf
+oid sha256:6c3011363c1e96100e580e8e942e83a5040524d32a7549efec4811d470255e73
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8d9dac6c5d8b3dbec92691bba6fb03d0b2f2d704fb193b9796315e27695e7797
-size 14180
+oid sha256:2185d1086b63cce3d1a18784aebd527c89731c1a6628c05136eec58027c1403a
+size 14244

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:82db5b009baee2fe47d69ea19cb9df28361cb0a44c82beb7be112c1f5681f538
+oid sha256:90da8f4c5f04c00fbf4044be4fa60fc55b99c69b8f4c1a2edff86d607394700a
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6cc08d586e3e1da1bc46b90ce3e134a3123c7b4622e7f31b236cfd3059766c1a
+oid sha256:53a2ad16e207bab6529f167d99325289b0e08bae4d436c0a57e9385c3e463e51
 size 1064

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 20250,
-  "best_metric": 0.7780515551567078,
-  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-20000",
-  "epoch": 1.576801784478117,
+  "best_global_step": 21000,
+  "best_metric": 0.7739425301551819,
+  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-21000",
+  "epoch": 1.6152603645873396,
   "eval_steps": 250,
-  "global_step": 20500,
+  "global_step": 21000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -3534,6 +3534,92 @@
       "eval_samples_per_second": 55.609,
       "eval_steps_per_second": 13.902,
       "step": 20500
+    },
+    {
+      "epoch": 1.5806476424890392,
+      "grad_norm": 0.781563401222229,
+      "learning_rate": 9.589902085551775e-05,
+      "loss": 0.7251,
+      "step": 20550
+    },
+    {
+      "epoch": 1.5844935004999616,
+      "grad_norm": 0.8129003047943115,
+      "learning_rate": 9.563930083370128e-05,
+      "loss": 0.7783,
+      "step": 20600
+    },
+    {
+      "epoch": 1.5883393585108838,
+      "grad_norm": 0.7955138087272644,
+      "learning_rate": 9.537958081188479e-05,
+      "loss": 0.7692,
+      "step": 20650
+    },
+    {
+      "epoch": 1.592185216521806,
+      "grad_norm": 0.7752518653869629,
+      "learning_rate": 9.511986079006832e-05,
+      "loss": 0.7578,
+      "step": 20700
+    },
+    {
+      "epoch": 1.5960310745327284,
+      "grad_norm": 0.7433210611343384,
+      "learning_rate": 9.486014076825183e-05,
+      "loss": 0.7885,
+      "step": 20750
+    },
+    {
+      "epoch": 1.5960310745327284,
+      "eval_loss": 0.7757794260978699,
+      "eval_runtime": 17.9593,
+      "eval_samples_per_second": 55.682,
+      "eval_steps_per_second": 13.92,
+      "step": 20750
+    },
+    {
+      "epoch": 1.5998769325436504,
+      "grad_norm": 0.7218450903892517,
+      "learning_rate": 9.460042074643534e-05,
+      "loss": 0.8179,
+      "step": 20800
+    },
+    {
+      "epoch": 1.6037227905545728,
+      "grad_norm": 0.8611409664154053,
+      "learning_rate": 9.434070072461887e-05,
+      "loss": 0.8166,
+      "step": 20850
+    },
+    {
+      "epoch": 1.607568648565495,
+      "grad_norm": 0.8060470223426819,
+      "learning_rate": 9.408098070280238e-05,
+      "loss": 0.7831,
+      "step": 20900
+    },
+    {
+      "epoch": 1.6114145065764172,
+      "grad_norm": 0.9832955002784729,
+      "learning_rate": 9.382126068098591e-05,
+      "loss": 0.7679,
+      "step": 20950
+    },
+    {
+      "epoch": 1.6152603645873396,
+      "grad_norm": 0.7749195098876953,
+      "learning_rate": 9.356154065916942e-05,
+      "loss": 0.8001,
+      "step": 21000
+    },
+    {
+      "epoch": 1.6152603645873396,
+      "eval_loss": 0.7739425301551819,
+      "eval_runtime": 17.9752,
+      "eval_samples_per_second": 55.632,
+      "eval_steps_per_second": 13.908,
+      "step": 21000
     }
   ],
   "logging_steps": 50,