Saving weights and logs of step 15000

events.out.tfevents.1625843003.t1v-n-528d9406-w-0.250031.3.v2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3d95b12f427bf583b9067304b1335be00b71a33d36f953dad1f1e647a8f0526c
-size 1470349
+oid sha256:70d0348a3f3f5ea1e7a29c686369e14227778a5eedb6076a74980606d2f40bda
+size 2205560

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e12359d533ceae1fa223871492c62e14d0c586867a48c6ef7a6a2fc30357060a
+oid sha256:25e72a24af671ec402054b28c4d8d88e526ac5ebfd84c4270f8c8ff64b012c0b
 size 1419302302

wandb/run-20210709_144100-2k1kyrq2/files/output.log

@@ -424,3 +424,5 @@ loading file None
 Step... (5000 | Loss: 3.704026460647583, Learning Rate: 0.00238410709425807)
 Step... (5000 | Eval Loss: 3.6012017726898193 | Eval Perplexity: 36.64224364155734)
 Step... (10000 | Loss: 3.3888800144195557, Learning Rate: 0.0023642408195883036)
+Step... (10000 | Eval Loss: 3.424499034881592 | Eval Perplexity: 30.707257732613787)
+Step... (15000 | Loss: 3.297701358795166, Learning Rate: 0.002344374777749181)

wandb/run-20210709_144100-2k1kyrq2/files/wandb-summary.json

@@ -1 +1 @@
-{"global_step": 10000, "_timestamp": 1625854463.854469, "train_time": 17070.388671875, "train_learning_rate": 0.0023642408195883036, "_step": 19994, "train_loss": 3.383782386779785, "eval_loss": 3.6012017726898193, "eval_perplexity": 36.642242431640625}
+{"global_step": 15000, "_timestamp": 1625860279.738529, "train_time": 34334.51171875, "train_learning_rate": 0.002344374777749181, "_step": 29991, "train_loss": 3.3004040718078613, "eval_loss": 3.424499034881592, "eval_perplexity": 30.707258224487305}