Training in progress, step 21500, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3eb66659915116c9f52f8b7d65982f16dfe01a2f397d774e7e2e124b2fee93e6
+oid sha256:abe147d1fd4bc110db7089e008a574b65f554aaa98420567a0bc2b63656bd9c1
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:094124f70673a425259be33a9a30631e505be7e6b457e57a5337a37485da2199
+oid sha256:a3abcac0bf194b5337811637955f6e6b362380aa1f532202481201a600657324
 size 1736

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6c3011363c1e96100e580e8e942e83a5040524d32a7549efec4811d470255e73
+oid sha256:1d206c147f3ebf5a17a6c44dea5417a1ae1dd50a6fd22e30c6aea36175324b86
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2185d1086b63cce3d1a18784aebd527c89731c1a6628c05136eec58027c1403a
-size 14244
+oid sha256:4a5d062bb7d441f36a44929472f10d0b795d7b5f5a9dcd38edd739ffa77a6d36
+size 14180

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90da8f4c5f04c00fbf4044be4fa60fc55b99c69b8f4c1a2edff86d607394700a
+oid sha256:e00a972afc33ed5290e59d706523471e5c191feb09759aef8aeb13be22ede6e8
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:53a2ad16e207bab6529f167d99325289b0e08bae4d436c0a57e9385c3e463e51
+oid sha256:afb6a0769e42f6963785d597ff6271423cd7dd782771f3b3c0a05c348f1cd005
 size 1064

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 21000,
-  "best_metric": 0.7739425301551819,
-  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-21000",
-  "epoch": 1.6152603645873396,
+  "best_global_step": 21500,
+  "best_metric": 0.77164626121521,
+  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-21500",
+  "epoch": 1.6537189446965619,
   "eval_steps": 250,
-  "global_step": 21000,
+  "global_step": 21500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -3620,6 +3620,92 @@
       "eval_samples_per_second": 55.632,
       "eval_steps_per_second": 13.908,
       "step": 21000
+    },
+    {
+      "epoch": 1.6191062225982615,
+      "grad_norm": 0.8098833560943604,
+      "learning_rate": 9.330182063735295e-05,
+      "loss": 0.7673,
+      "step": 21050
+    },
+    {
+      "epoch": 1.622952080609184,
+      "grad_norm": 0.7867510318756104,
+      "learning_rate": 9.304210061553646e-05,
+      "loss": 0.7619,
+      "step": 21100
+    },
+    {
+      "epoch": 1.6267979386201061,
+      "grad_norm": 1.2399013042449951,
+      "learning_rate": 9.278238059371997e-05,
+      "loss": 0.8664,
+      "step": 21150
+    },
+    {
+      "epoch": 1.6306437966310283,
+      "grad_norm": 0.6171821355819702,
+      "learning_rate": 9.25226605719035e-05,
+      "loss": 0.8543,
+      "step": 21200
+    },
+    {
+      "epoch": 1.6344896546419507,
+      "grad_norm": 0.7197456955909729,
+      "learning_rate": 9.226294055008701e-05,
+      "loss": 0.7989,
+      "step": 21250
+    },
+    {
+      "epoch": 1.6344896546419507,
+      "eval_loss": 0.7736220955848694,
+      "eval_runtime": 18.0274,
+      "eval_samples_per_second": 55.471,
+      "eval_steps_per_second": 13.868,
+      "step": 21250
+    },
+    {
+      "epoch": 1.6383355126528727,
+      "grad_norm": 0.6752798557281494,
+      "learning_rate": 9.200322052827052e-05,
+      "loss": 0.7925,
+      "step": 21300
+    },
+    {
+      "epoch": 1.642181370663795,
+      "grad_norm": 0.7389090061187744,
+      "learning_rate": 9.174350050645405e-05,
+      "loss": 0.762,
+      "step": 21350
+    },
+    {
+      "epoch": 1.6460272286747173,
+      "grad_norm": 0.7688984870910645,
+      "learning_rate": 9.148378048463756e-05,
+      "loss": 0.771,
+      "step": 21400
+    },
+    {
+      "epoch": 1.6498730866856395,
+      "grad_norm": 0.7231914401054382,
+      "learning_rate": 9.122406046282108e-05,
+      "loss": 0.8244,
+      "step": 21450
+    },
+    {
+      "epoch": 1.6537189446965619,
+      "grad_norm": 0.786527156829834,
+      "learning_rate": 9.09643404410046e-05,
+      "loss": 0.7738,
+      "step": 21500
+    },
+    {
+      "epoch": 1.6537189446965619,
+      "eval_loss": 0.77164626121521,
+      "eval_runtime": 17.9322,
+      "eval_samples_per_second": 55.765,
+      "eval_steps_per_second": 13.941,
+      "step": 21500
     }
   ],
   "logging_steps": 50,