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:e938b4a2a45d05421534909af0bf009dedebb9aea74565d4ae4302594883dc8f
+oid sha256:14524230b3f005df78a646dd1b27d41289ac1f66834efa675f202b683ceaaac3
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7455f8d2c994f64aaeae6d25909ff5a023afe0e1744ef0734509e43b10f56157
+oid sha256:1066bdd6c1caadf20ebc14f9d07b52f0568e1887cdd278685cc8d7c58198d96d
 size 1736

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6cd326cd8b59bceb89f65022b035ca69baf384342ec2bf39dcc4ad1f53650ae
+oid sha256:272013b98c63908bfc45574a8fa9d70a73d1d3eb790b0f8a303e3a15b837cf4d
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:53b32835bd0fc28ffeb2c034cbe6942df7d91018aa69d3614686950a58fc9920
+oid sha256:40f79270230c464fa914f6a807febad7776661ad4bd2b9ac4902f678e7bb8c9e
 size 14244

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2dda7fe4be06e7b2f28f9fb6d77a55fd50f883f30376b3be3fec2001c9515fe5
+oid sha256:9010363377eeaf4fa17beb36b7862690464999e47731a3fe7bab298c1d354d96
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:24a9ece5494351db57cf0201c68e80d1bab27482ad54e226480b7b466bc789e5
+oid sha256:89b75f3e11e5fda95765f8aa3a44fc74ade3f238246295323c8bfd8ff24e728a
 size 1064

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 19500,
-  "best_metric": 0.7843549847602844,
-  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-19500",
-  "epoch": 1.4998846242596724,
+  "best_global_step": 20000,
+  "best_metric": 0.7793118953704834,
+  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-20000",
+  "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": 55.807,
       "eval_steps_per_second": 13.952,
       "step": 19500
+    },
+    {
+      "epoch": 1.5037304822705946,
+      "grad_norm": 0.5136408805847168,
+      "learning_rate": 0.0001010934212918474,
+      "loss": 0.7533,
+      "step": 19550
+    },
+    {
+      "epoch": 1.5075763402815168,
+      "grad_norm": 0.8944385647773743,
+      "learning_rate": 0.00010083370127003091,
+      "loss": 0.8483,
+      "step": 19600
+    },
+    {
+      "epoch": 1.511422198292439,
+      "grad_norm": 0.9759209752082825,
+      "learning_rate": 0.00010057398124821442,
+      "loss": 0.8338,
+      "step": 19650
+    },
+    {
+      "epoch": 1.5152680563033614,
+      "grad_norm": 0.7823068499565125,
+      "learning_rate": 0.00010031426122639795,
+      "loss": 0.7739,
+      "step": 19700
+    },
+    {
+      "epoch": 1.5191139143142836,
+      "grad_norm": 0.9423583745956421,
+      "learning_rate": 0.00010005454120458146,
+      "loss": 0.8092,
+      "step": 19750
+    },
+    {
+      "epoch": 1.5191139143142836,
+      "eval_loss": 0.782052218914032,
+      "eval_runtime": 17.8818,
+      "eval_samples_per_second": 55.923,
+      "eval_steps_per_second": 13.981,
+      "step": 19750
+    },
+    {
+      "epoch": 1.5229597723252057,
+      "grad_norm": 1.067384123802185,
+      "learning_rate": 9.979482118276499e-05,
+      "loss": 0.7581,
+      "step": 19800
+    },
+    {
+      "epoch": 1.5268056303361282,
+      "grad_norm": 1.0136195421218872,
+      "learning_rate": 9.95351011609485e-05,
+      "loss": 0.7923,
+      "step": 19850
+    },
+    {
+      "epoch": 1.5306514883470501,
+      "grad_norm": 0.8290442824363708,
+      "learning_rate": 9.927538113913202e-05,
+      "loss": 0.7991,
+      "step": 19900
+    },
+    {
+      "epoch": 1.5344973463579725,
+      "grad_norm": 0.8200196623802185,
+      "learning_rate": 9.901566111731554e-05,
+      "loss": 0.7382,
+      "step": 19950
+    },
+    {
+      "epoch": 1.5383432043688947,
+      "grad_norm": 0.7905208468437195,
+      "learning_rate": 9.875594109549905e-05,
+      "loss": 0.7615,
+      "step": 20000
+    },
+    {
+      "epoch": 1.5383432043688947,
+      "eval_loss": 0.7793118953704834,
+      "eval_runtime": 17.8391,
+      "eval_samples_per_second": 56.057,
+      "eval_steps_per_second": 14.014,
+      "step": 20000
     }
   ],
   "logging_steps": 50,