Training in progress, epoch 0, checkpoint

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:538342ac688fb0a0d86aaedd7330737ae00e80eb8e093f459c58ce63d811fc33
+oid sha256:1f801c41a48a686d1e8c361a6732caaf9bafd7adba82937dd2b58b735832368a
 size 1256

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2ba8427ac6eea57ec5734699585fe0dd282eb1503ee998fdc4232e54bb7d6354
+oid sha256:ae28f9b2f5015b95ad2f650d54d1979a68d40e31cc70d2fd5c54ddf3ae4a6519
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0bc60bce785072809819f1c1fc75a413381f48d7fd0b438c4400860ca72a7129
+oid sha256:7c70b4498d2f1f0f22fc7b37a8cc650fd9782887b3bd68443093cb35a029f0af
 size 1064

last-checkpoint/trainer_state.json

@@ -1,188 +1,25 @@
 {
-  "best_metric": 2.2744693756103516,
-  "best_model_checkpoint": "dq158/pingusPongus/checkpoint-6323",
-  "epoch": 2.0,
+  "best_metric": 2.096945285797119,
+  "best_model_checkpoint": "dq158/pingusPongus/checkpoint-790",
+  "epoch": 0.9993674889310563,
   "eval_steps": 500,
-  "global_step": 12646,
+  "global_step": 790,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
-    {
-      "epoch": 0.08,
-      "learning_rate": 8.990024286845271e-05,
-      "loss": 2.3525,
-      "step": 500
-    },
-    {
-      "epoch": 0.16,
-      "learning_rate": 8.949573742576993e-05,
-      "loss": 2.3648,
-      "step": 1000
-    },
-    {
-      "epoch": 0.24,
-      "learning_rate": 8.878304835100916e-05,
-      "loss": 2.3927,
-      "step": 1500
-    },
-    {
-      "epoch": 0.32,
-      "learning_rate": 8.776711182728166e-05,
-      "loss": 2.3619,
-      "step": 2000
-    },
-    {
-      "epoch": 0.4,
-      "learning_rate": 8.645496437147547e-05,
-      "loss": 2.3352,
-      "step": 2500
-    },
-    {
-      "epoch": 0.47,
-      "learning_rate": 8.485569409836621e-05,
-      "loss": 2.3968,
-      "step": 3000
-    },
-    {
-      "epoch": 0.55,
-      "learning_rate": 8.298037777507714e-05,
-      "loss": 2.3855,
-      "step": 3500
-    },
     {
       "epoch": 0.63,
-      "learning_rate": 8.084200410185778e-05,
-      "loss": 2.3609,
-      "step": 4000
-    },
-    {
-      "epoch": 0.71,
-      "learning_rate": 7.845538375054953e-05,
-      "loss": 2.3766,
-      "step": 4500
-    },
-    {
-      "epoch": 0.79,
-      "learning_rate": 7.583704678382374e-05,
-      "loss": 2.3645,
-      "step": 5000
-    },
-    {
-      "epoch": 0.87,
-      "learning_rate": 7.300512816568131e-05,
-      "loss": 2.3709,
-      "step": 5500
-    },
-    {
-      "epoch": 0.95,
-      "learning_rate": 6.997924215618355e-05,
-      "loss": 2.3794,
-      "step": 6000
+      "learning_rate": 4.6266135493489015e-05,
+      "loss": 2.2387,
+      "step": 500
     },
     {
       "epoch": 1.0,
       "eval_bleu": 1.0,
       "eval_brevity_penalty": 1.0,
       "eval_length_ratio": 1.0,
-      "eval_loss": 2.2744693756103516,
-      "eval_precisions": [
-        1.0,
-        1.0,
-        1.0,
-        1.0
-      ],
-      "eval_reference_length": 1439232,
-      "eval_runtime": 1571.1114,
-      "eval_samples_per_second": 1.789,
-      "eval_steps_per_second": 0.895,
-      "eval_translation_length": 1439232,
-      "step": 6323
-    },
-    {
-      "epoch": 1.03,
-      "learning_rate": 6.678034646037393e-05,
-      "loss": 2.3216,
-      "step": 6500
-    },
-    {
-      "epoch": 1.11,
-      "learning_rate": 6.343059707231406e-05,
-      "loss": 2.3675,
-      "step": 7000
-    },
-    {
-      "epoch": 1.19,
-      "learning_rate": 5.995319481960425e-05,
-      "loss": 2.3563,
-      "step": 7500
-    },
-    {
-      "epoch": 1.27,
-      "learning_rate": 5.6372224671242366e-05,
-      "loss": 2.4016,
-      "step": 8000
-    },
-    {
-      "epoch": 1.34,
-      "learning_rate": 5.2712488921797064e-05,
-      "loss": 2.3769,
-      "step": 8500
-    },
-    {
-      "epoch": 1.42,
-      "learning_rate": 4.899933540728503e-05,
-      "loss": 2.3531,
-      "step": 9000
-    },
-    {
-      "epoch": 1.5,
-      "learning_rate": 4.5258481942552583e-05,
-      "loss": 2.3503,
-      "step": 9500
-    },
-    {
-      "epoch": 1.58,
-      "learning_rate": 4.151583819613269e-05,
-      "loss": 2.3508,
-      "step": 10000
-    },
-    {
-      "epoch": 1.66,
-      "learning_rate": 3.779732623629695e-05,
-      "loss": 2.3509,
-      "step": 10500
-    },
-    {
-      "epoch": 1.74,
-      "learning_rate": 3.412870099122515e-05,
-      "loss": 2.3825,
-      "step": 11000
-    },
-    {
-      "epoch": 1.82,
-      "learning_rate": 3.0535371866810265e-05,
-      "loss": 2.3884,
-      "step": 11500
-    },
-    {
-      "epoch": 1.9,
-      "learning_rate": 2.704222675759866e-05,
-      "loss": 2.3531,
-      "step": 12000
-    },
-    {
-      "epoch": 1.98,
-      "learning_rate": 2.367345966978985e-05,
-      "loss": 2.3813,
-      "step": 12500
-    },
-    {
-      "epoch": 2.0,
-      "eval_bleu": 1.0,
-      "eval_brevity_penalty": 1.0,
-      "eval_length_ratio": 1.0,
-      "eval_loss": 2.2744693756103516,
+      "eval_loss": 2.096945285797119,
       "eval_precisions": [
         1.0,
         1.0,
@@ -190,18 +27,18 @@
         1.0
       ],
       "eval_reference_length": 1439232,
-      "eval_runtime": 1567.8247,
-      "eval_samples_per_second": 1.793,
-      "eval_steps_per_second": 0.897,
+      "eval_runtime": 879.4262,
+      "eval_samples_per_second": 3.196,
+      "eval_steps_per_second": 0.2,
       "eval_translation_length": 1439232,
-      "step": 12646
+      "step": 790
     }
   ],
   "logging_steps": 500,
-  "max_steps": 18969,
+  "max_steps": 2370,
   "num_train_epochs": 3,
   "save_steps": 500,
-  "total_flos": 4.3332645393373594e+17,
+  "total_flos": 2.1666322696686797e+17,
   "trial_name": null,
   "trial_params": null
 }

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7ef329a26ff1c545c7abfacd51a0506adc651e12a2a47e71cffbb6c7a728715a
+oid sha256:b7694805b29bfb8491caf50687f3617fba8e8b17c948a4fabc0da0476e235ec2
 size 4664