Training in progress, step 2000, checkpoint

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:50f4e73a7428807d0545ea63e75cb228396f0cccd6fbcd5ae2eaf2557ecae994
-size 1417163
+oid sha256:1f907b0bed5e2f8cf858db047aa22f79320ace855462d3e01515d8384d133c06
+size 1976267

last-checkpoint/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fc861cf9df458b75f1604d94b61feea15513b6ffde8bb6828abb654c32ba1ea8
-size 292928643
+oid sha256:5ed1d8ea23bc21d353cc09795bf8da7e15e7897910a93e297620cb29f241e863
+size 741336195

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87cfd50ff3a42065a545f46e5e0d163da023498a41ad02da368f633d32b47d2c
+oid sha256:1598f69323c5211489e239f2a4a33d79e9afe5826ee72973f1d7e1a5afa7e09a
 size 14455

last-checkpoint/trainer_state.json

@@ -11,160 +11,160 @@
   "log_history": [
     {
       "epoch": 0.002,
-      "grad_norm": 14.826459884643555,
+      "grad_norm": 12.660503387451172,
       "learning_rate": 7.920000000000001e-07,
-      "loss": 10.879287109375,
+      "loss": 10.891383056640626,
       "step": 100
     },
     {
       "epoch": 0.004,
-      "grad_norm": 17.237497329711914,
+      "grad_norm": 7.186215400695801,
       "learning_rate": 1.5920000000000002e-06,
-      "loss": 10.74796630859375,
+      "loss": 10.36431640625,
       "step": 200
     },
     {
       "epoch": 0.006,
-      "grad_norm": 5.412790298461914,
+      "grad_norm": 3.448742389678955,
       "learning_rate": 2.392e-06,
-      "loss": 10.419390869140624,
+      "loss": 9.655166625976562,
       "step": 300
     },
     {
       "epoch": 0.008,
-      "grad_norm": 3.887077569961548,
+      "grad_norm": 3.1286885738372803,
       "learning_rate": 3.192e-06,
-      "loss": 10.083310546875,
+      "loss": 9.2309130859375,
       "step": 400
     },
     {
       "epoch": 0.01,
-      "grad_norm": 2.948329448699951,
+      "grad_norm": 2.5423293113708496,
       "learning_rate": 3.992e-06,
-      "loss": 9.929519653320312,
+      "loss": 9.036063232421874,
       "step": 500
     },
     {
       "epoch": 0.012,
-      "grad_norm": 3.4345157146453857,
+      "grad_norm": 3.8436737060546875,
       "learning_rate": 4.792000000000001e-06,
-      "loss": 9.777420654296876,
+      "loss": 8.7992333984375,
       "step": 600
     },
     {
       "epoch": 0.014,
-      "grad_norm": 4.727835655212402,
+      "grad_norm": 3.467728614807129,
       "learning_rate": 5.592000000000001e-06,
-      "loss": 9.646694946289063,
+      "loss": 8.584224853515625,
       "step": 700
     },
     {
       "epoch": 0.016,
-      "grad_norm": 2.9691734313964844,
+      "grad_norm": 2.465226411819458,
       "learning_rate": 6.392000000000001e-06,
-      "loss": 9.485365600585938,
+      "loss": 8.331553955078125,
       "step": 800
     },
     {
       "epoch": 0.018,
-      "grad_norm": 8.859844207763672,
+      "grad_norm": 8.884796142578125,
       "learning_rate": 7.192e-06,
-      "loss": 9.32474365234375,
+      "loss": 8.120615234375,
       "step": 900
     },
     {
       "epoch": 0.02,
-      "grad_norm": 4.142549514770508,
+      "grad_norm": 3.531571388244629,
       "learning_rate": 7.992e-06,
-      "loss": 9.14423828125,
+      "loss": 7.909391479492188,
       "step": 1000
     },
     {
       "epoch": 0.02,
-      "eval_accuracy": 0.09021526418786693,
-      "eval_loss": 9.03640079498291,
-      "eval_runtime": 20.6098,
-      "eval_samples_per_second": 48.521,
-      "eval_steps_per_second": 6.065,
+      "eval_accuracy": 0.10772407045009785,
+      "eval_loss": 7.779632091522217,
+      "eval_runtime": 20.2829,
+      "eval_samples_per_second": 49.303,
+      "eval_steps_per_second": 6.163,
       "step": 1000
     },
     {
       "epoch": 0.022,
-      "grad_norm": 6.929040431976318,
+      "grad_norm": 3.3351476192474365,
       "learning_rate": 8.792e-06,
-      "loss": 8.938278198242188,
+      "loss": 7.69487060546875,
       "step": 1100
     },
     {
       "epoch": 0.024,
-      "grad_norm": 3.2667243480682373,
+      "grad_norm": 6.731584072113037,
       "learning_rate": 9.592e-06,
-      "loss": 8.691138916015625,
+      "loss": 7.469768676757813,
       "step": 1200
     },
     {
       "epoch": 0.026,
-      "grad_norm": 3.8360178470611572,
+      "grad_norm": 2.172372817993164,
       "learning_rate": 1.0392e-05,
-      "loss": 8.525877685546876,
+      "loss": 7.39029052734375,
       "step": 1300
     },
     {
       "epoch": 0.028,
-      "grad_norm": 7.723220348358154,
+      "grad_norm": 2.362459421157837,
       "learning_rate": 1.1192e-05,
-      "loss": 8.324917602539063,
+      "loss": 7.290626220703125,
       "step": 1400
     },
     {
       "epoch": 0.03,
-      "grad_norm": 4.095365047454834,
+      "grad_norm": 2.927532196044922,
       "learning_rate": 1.1992000000000001e-05,
-      "loss": 8.129039916992188,
+      "loss": 7.201221313476562,
       "step": 1500
     },
     {
       "epoch": 0.032,
-      "grad_norm": 8.2203950881958,
+      "grad_norm": 3.7425637245178223,
       "learning_rate": 1.2792e-05,
-      "loss": 7.922721557617187,
+      "loss": 7.103947143554688,
       "step": 1600
     },
     {
       "epoch": 0.034,
-      "grad_norm": 3.7490456104278564,
+      "grad_norm": 3.8855597972869873,
       "learning_rate": 1.3592000000000001e-05,
-      "loss": 7.799203491210937,
+      "loss": 7.126583862304687,
       "step": 1700
     },
     {
       "epoch": 0.036,
-      "grad_norm": 4.54710054397583,
+      "grad_norm": 3.833674669265747,
       "learning_rate": 1.4392000000000002e-05,
-      "loss": 7.621753540039062,
+      "loss": 7.025943603515625,
       "step": 1800
     },
     {
       "epoch": 0.038,
-      "grad_norm": 2.902744770050049,
+      "grad_norm": 3.688350200653076,
       "learning_rate": 1.5192000000000003e-05,
-      "loss": 7.5074224853515625,
+      "loss": 6.976273803710938,
       "step": 1900
     },
     {
       "epoch": 0.04,
-      "grad_norm": 6.133886337280273,
+      "grad_norm": 3.1060683727264404,
       "learning_rate": 1.5992000000000002e-05,
-      "loss": 7.375697631835937,
+      "loss": 6.903162841796875,
       "step": 2000
     },
     {
       "epoch": 0.04,
-      "eval_accuracy": 0.1294050880626223,
-      "eval_loss": 7.383677959442139,
-      "eval_runtime": 7.6198,
-      "eval_samples_per_second": 131.237,
-      "eval_steps_per_second": 16.405,
+      "eval_accuracy": 0.14267514677103718,
+      "eval_loss": 6.938798904418945,
+      "eval_runtime": 8.7308,
+      "eval_samples_per_second": 114.538,
+      "eval_steps_per_second": 14.317,
       "step": 2000
     }
   ],