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:f82c93836bc692e34902f10bc204524481cd7a576e6d3e28b3bb7e2d682f323f
+oid sha256:d257112e537fe70c79019496ce8a965aa34d1fb175b35addc498d01e38be6234
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6e8703e2ac4edb79cf8f1764c40656f93e9be69ded66eb4cce14bbce6431a5e0
+oid sha256:25ff7761b7be0ad2ef125a7c44f3a00d691af1455b26167da65e53a9f3d4a4e1
 size 1544

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:411bf70b4f8c28dcde0ada0a6ab13ec881cbbf19e85c2796efaba99e8b13583e
+oid sha256:f34cf1c465a3f22b58ed5814e553f1f151f9d6086122d8e635e6c9a1efd7a02c
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5dbffb5b9d940a96eb51dbe7a35860718ecfeca99437ad8b1b71dbafaefdd573
+oid sha256:4c5132bfe30a63a6382081eeb67e311e07e2d98916afbc38ab8f8ba0c010b059
 size 14244

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:384ba37c8844299a29c3c998811c9d47320ed4f7666069a4fb9a63d0dddfe44f
+oid sha256:279d58c28b18870f94d45189b49d69a8021c82d9a61e08eb069f8a47562cd494
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c62f9fce1be162dfb9c17d3cf04dc0505b42c85ac9a955fe99836e3dd2e88351
+oid sha256:12146cd21738b806f2d3b5e8ca727dc812aefa4523582c6a525173e0b62cdd2a
 size 1064

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 20250,
-  "best_metric": 1.4724150896072388,
-  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-15500",
-  "epoch": 1.576801784478117,
+  "best_global_step": 21000,
+  "best_metric": 1.4690666198730469,
+  "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": 56.061,
       "eval_steps_per_second": 14.015,
       "step": 20500
+    },
+    {
+      "epoch": 1.5806476424890392,
+      "grad_norm": 1.430114507675171,
+      "learning_rate": 9.590940965639041e-05,
+      "loss": 1.3714,
+      "step": 20550
+    },
+    {
+      "epoch": 1.5844935004999616,
+      "grad_norm": 1.7613717317581177,
+      "learning_rate": 9.564968963457394e-05,
+      "loss": 1.4285,
+      "step": 20600
+    },
+    {
+      "epoch": 1.5883393585108838,
+      "grad_norm": 1.3678529262542725,
+      "learning_rate": 9.538996961275745e-05,
+      "loss": 1.4463,
+      "step": 20650
+    },
+    {
+      "epoch": 1.592185216521806,
+      "grad_norm": 3.1135504245758057,
+      "learning_rate": 9.513024959094097e-05,
+      "loss": 1.4287,
+      "step": 20700
+    },
+    {
+      "epoch": 1.5960310745327284,
+      "grad_norm": 1.7866570949554443,
+      "learning_rate": 9.487052956912449e-05,
+      "loss": 1.4515,
+      "step": 20750
+    },
+    {
+      "epoch": 1.5960310745327284,
+      "eval_loss": 1.4817472696304321,
+      "eval_runtime": 17.9267,
+      "eval_samples_per_second": 55.783,
+      "eval_steps_per_second": 13.946,
+      "step": 20750
+    },
+    {
+      "epoch": 1.5998769325436504,
+      "grad_norm": 1.2406786680221558,
+      "learning_rate": 9.4610809547308e-05,
+      "loss": 1.5364,
+      "step": 20800
+    },
+    {
+      "epoch": 1.6037227905545728,
+      "grad_norm": 1.9183951616287231,
+      "learning_rate": 9.435108952549152e-05,
+      "loss": 1.5169,
+      "step": 20850
+    },
+    {
+      "epoch": 1.607568648565495,
+      "grad_norm": 1.852089524269104,
+      "learning_rate": 9.409136950367504e-05,
+      "loss": 1.4668,
+      "step": 20900
+    },
+    {
+      "epoch": 1.6114145065764172,
+      "grad_norm": 2.3430335521698,
+      "learning_rate": 9.383164948185856e-05,
+      "loss": 1.4297,
+      "step": 20950
+    },
+    {
+      "epoch": 1.6152603645873396,
+      "grad_norm": 1.9646743535995483,
+      "learning_rate": 9.357192946004207e-05,
+      "loss": 1.5143,
+      "step": 21000
+    },
+    {
+      "epoch": 1.6152603645873396,
+      "eval_loss": 1.4690666198730469,
+      "eval_runtime": 17.7563,
+      "eval_samples_per_second": 56.318,
+      "eval_steps_per_second": 14.079,
+      "step": 21000
     }
   ],
   "logging_steps": 50,