Training in progress, step 18000, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c504a27eb69195f0629a47f9e0bf192539abb3c2443da0cc76cbb0d75cfd5228
+oid sha256:ab1600c15a37ec6fb41eca032481fb072aba7998a063361aac9692ac6a9ef216
 size 36730224

last-checkpoint/ar_diffusion_info.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f425ddc672454ec318d3a477b01d57b3febdbf4bb7f94794d250d01447f38c3
+oid sha256:16cac24c8d65482bd4344116193b8cb0973617948ed3db3f9320f39d3b079b5f
 size 1736

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c2c9d67d788a0699ac5c4a4b6a7c0a6433a765152542a0d4c63e46c093056b65
+oid sha256:245296adc55ccbb77650410469534fb289b0cd17f03dec26c0aee8f21a8e2682
 size 73588346

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a8ffe3d878b186570864a01f18402ac15fa8e666ba6230c15554b4863a170b12
+oid sha256:69e9bd7ace00dc839f848c20ae8e1e8fe1387810881778a2284fcf0ac2ad2525
 size 14244

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e4317872e22f122cc23eecd5b16d2ae8cc3637fa26840af5163be9ef4599ef9
+oid sha256:ac398991953ddd2d4574773fb96cf571b99bfe36340ab697de86477c38530f18
 size 988

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1f03e29ee3758301caaed8e1dd1dbe2a845f2652e79a6f8508f414e5c4c84bdf
+oid sha256:8d68c6e80dd0b62a913dae04c12415a6a255632f5af4175f8cdb6a83a0bbfe28
 size 1064

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 17250,
-  "best_metric": 0.7951701879501343,
-  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-17000",
-  "epoch": 1.346050303822783,
+  "best_global_step": 18000,
+  "best_metric": 0.7882509231567383,
+  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-18000",
+  "epoch": 1.3845088839320052,
   "eval_steps": 250,
-  "global_step": 17500,
+  "global_step": 18000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -3018,6 +3018,92 @@
       "eval_samples_per_second": 55.577,
       "eval_steps_per_second": 13.894,
       "step": 17500
+    },
+    {
+      "epoch": 1.349896161833705,
+      "grad_norm": 0.8688609600067139,
+      "learning_rate": 0.00011148222216450668,
+      "loss": 0.7882,
+      "step": 17550
+    },
+    {
+      "epoch": 1.3537420198446273,
+      "grad_norm": 0.7067499160766602,
+      "learning_rate": 0.0001112225021426902,
+      "loss": 0.8193,
+      "step": 17600
+    },
+    {
+      "epoch": 1.3575878778555497,
+      "grad_norm": 0.8119627833366394,
+      "learning_rate": 0.0001109627821208737,
+      "loss": 0.7902,
+      "step": 17650
+    },
+    {
+      "epoch": 1.3614337358664719,
+      "grad_norm": 0.4667348265647888,
+      "learning_rate": 0.00011070306209905723,
+      "loss": 0.7595,
+      "step": 17700
+    },
+    {
+      "epoch": 1.365279593877394,
+      "grad_norm": 1.1219276189804077,
+      "learning_rate": 0.00011044334207724075,
+      "loss": 0.8049,
+      "step": 17750
+    },
+    {
+      "epoch": 1.365279593877394,
+      "eval_loss": 0.7947649955749512,
+      "eval_runtime": 17.9375,
+      "eval_samples_per_second": 55.749,
+      "eval_steps_per_second": 13.937,
+      "step": 17750
+    },
+    {
+      "epoch": 1.3691254518883162,
+      "grad_norm": 0.8262600302696228,
+      "learning_rate": 0.00011018362205542425,
+      "loss": 0.794,
+      "step": 17800
+    },
+    {
+      "epoch": 1.3729713098992384,
+      "grad_norm": 0.6738994121551514,
+      "learning_rate": 0.00010992390203360779,
+      "loss": 0.7749,
+      "step": 17850
+    },
+    {
+      "epoch": 1.3768171679101608,
+      "grad_norm": 0.40902426838874817,
+      "learning_rate": 0.0001096641820117913,
+      "loss": 0.8111,
+      "step": 17900
+    },
+    {
+      "epoch": 1.380663025921083,
+      "grad_norm": 0.7617566585540771,
+      "learning_rate": 0.00010940446198997481,
+      "loss": 0.7663,
+      "step": 17950
+    },
+    {
+      "epoch": 1.3845088839320052,
+      "grad_norm": 0.5269647836685181,
+      "learning_rate": 0.00010914474196815834,
+      "loss": 0.7913,
+      "step": 18000
+    },
+    {
+      "epoch": 1.3845088839320052,
+      "eval_loss": 0.7882509231567383,
+      "eval_runtime": 17.938,
+      "eval_samples_per_second": 55.748,
+      "eval_steps_per_second": 13.937,
+      "step": 18000
     }
   ],
   "logging_steps": 50,