Training in progress, epoch 5, checkpoint

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5dcdbfaa9be1baa667851c48d1ed84ed40b0a797ad5ee982b83bc3b53e757697
+oid sha256:a1bab3f6d145c2dc15b63d98c856ba61dd4136ea20b9b56c6ac8e84167271447
 size 14244

last-checkpoint/trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 8.740928649902344,
-  "best_model_checkpoint": "./results/checkpoint-18888",
+  "best_metric": 8.15411631266276,
+  "best_model_checkpoint": "./results/checkpoint-23610",
   "epoch": 5.0,
   "eval_steps": 500,
   "global_step": 23610,
@@ -1707,6 +1707,18 @@
       "learning_rate": 1.2706480304955527e-08,
       "loss": 8.4148,
       "step": 23600
+    },
+    {
+      "epoch": 5.0,
+      "eval_avg_mae": 8.15411631266276,
+      "eval_loss": 8.1541166305542,
+      "eval_mae_lex": 8.678616523742676,
+      "eval_mae_sem": 4.634727478027344,
+      "eval_mae_syn": 11.149005889892578,
+      "eval_runtime": 276.8272,
+      "eval_samples_per_second": 136.999,
+      "eval_steps_per_second": 4.284,
+      "step": 23610
     }
   ],
   "logging_steps": 100,