Trained model with more epochs.

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7821d0cbfa96cb9e3dc1a9d11c3798188549e0b60e98b2bf452a78a21fe9a701
+oid sha256:d8130dcd44bc26bc493210590186970609afda402c8a8c30d783b6c38ce3b369
 size 995605189

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87039797eeee31fa3a1d10c2d523177194c55dc59f61f4e681b323a3ff27e3ac
+oid sha256:e338a12b8bebe4e2678f3e2df6bb4349cf4c19e37d8577fff4193457b40c2a63
 size 497807197

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:46ce6d1e3d7df661bdec831f2ccc0a92148db2766e8bae2d96b45ecc70cfa48d
+oid sha256:3e8b05db3db8612d55f664401de73515ed39d7840e12323f51fcce11ead5d8ac
 size 14575

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8015d3b05a1d6e704db23f7bbf89c9b1f626d36f6bdc05c31adde13e38374441
+oid sha256:3da2be91dd0fb10ee5c14ceb065a3e14a48dc06697808fe40026425eebb46709
 size 627

trainer_state.json

@@ -1,324 +1,268 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 4.839419269687637,
-  "global_step": 11000,
+  "epoch": 7.915567282321899,
+  "global_step": 9000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.22,
-      "learning_rate": 2.5e-05,
-      "loss": 7.2786,
+      "epoch": 0.44,
+      "learning_rate": 5e-05,
+      "loss": 6.3746,
       "step": 500
     },
     {
-      "epoch": 0.22,
-      "eval_loss": 4.886257171630859,
-      "eval_runtime": 81.5985,
-      "eval_samples_per_second": 190.972,
-      "eval_steps_per_second": 11.936,
+      "epoch": 0.44,
+      "eval_loss": 4.111661911010742,
+      "eval_runtime": 79.373,
+      "eval_samples_per_second": 196.326,
+      "eval_steps_per_second": 6.136,
       "step": 500
     },
     {
-      "epoch": 0.44,
-      "learning_rate": 5e-05,
-      "loss": 4.3573,
+      "epoch": 0.88,
+      "learning_rate": 0.0001,
+      "loss": 3.696,
       "step": 1000
     },
     {
-      "epoch": 0.44,
-      "eval_loss": 3.8646771907806396,
-      "eval_runtime": 81.3815,
-      "eval_samples_per_second": 191.481,
-      "eval_steps_per_second": 11.968,
+      "epoch": 0.88,
+      "eval_loss": 3.2242989540100098,
+      "eval_runtime": 79.3675,
+      "eval_samples_per_second": 196.34,
+      "eval_steps_per_second": 6.136,
       "step": 1000
     },
     {
-      "epoch": 0.66,
-      "learning_rate": 4.758803666184274e-05,
-      "loss": 3.6885,
+      "epoch": 1.32,
+      "learning_rate": 9.382411067193676e-05,
+      "loss": 2.9752,
       "step": 1500
     },
     {
-      "epoch": 0.66,
-      "eval_loss": 3.40698504447937,
-      "eval_runtime": 81.4183,
-      "eval_samples_per_second": 191.394,
-      "eval_steps_per_second": 11.963,
+      "epoch": 1.32,
+      "eval_loss": 2.7193033695220947,
+      "eval_runtime": 79.2691,
+      "eval_samples_per_second": 196.584,
+      "eval_steps_per_second": 6.144,
       "step": 1500
     },
     {
-      "epoch": 0.88,
-      "learning_rate": 4.517607332368548e-05,
-      "loss": 3.3144,
+      "epoch": 1.76,
+      "learning_rate": 8.764822134387353e-05,
+      "loss": 2.6145,
       "step": 2000
     },
     {
-      "epoch": 0.88,
-      "eval_loss": 3.0980827808380127,
-      "eval_runtime": 81.5071,
-      "eval_samples_per_second": 191.186,
-      "eval_steps_per_second": 11.95,
+      "epoch": 1.76,
+      "eval_loss": 2.45035457611084,
+      "eval_runtime": 79.2794,
+      "eval_samples_per_second": 196.558,
+      "eval_steps_per_second": 6.143,
       "step": 2000
     },
     {
-      "epoch": 1.1,
-      "learning_rate": 4.2764109985528225e-05,
-      "loss": 3.0218,
+      "epoch": 2.2,
+      "learning_rate": 8.147233201581028e-05,
+      "loss": 2.3557,
       "step": 2500
     },
     {
-      "epoch": 1.1,
-      "eval_loss": 2.8812012672424316,
-      "eval_runtime": 81.6765,
-      "eval_samples_per_second": 190.789,
-      "eval_steps_per_second": 11.925,
+      "epoch": 2.2,
+      "eval_loss": 2.2841453552246094,
+      "eval_runtime": 79.2481,
+      "eval_samples_per_second": 196.636,
+      "eval_steps_per_second": 6.145,
       "step": 2500
     },
     {
-      "epoch": 1.32,
-      "learning_rate": 4.0352146647370966e-05,
-      "loss": 2.832,
+      "epoch": 2.64,
+      "learning_rate": 7.529644268774703e-05,
+      "loss": 2.2026,
       "step": 3000
     },
     {
-      "epoch": 1.32,
-      "eval_loss": 2.730829954147339,
-      "eval_runtime": 81.679,
-      "eval_samples_per_second": 190.783,
-      "eval_steps_per_second": 11.925,
+      "epoch": 2.64,
+      "eval_loss": 2.173915386199951,
+      "eval_runtime": 79.3117,
+      "eval_samples_per_second": 196.478,
+      "eval_steps_per_second": 6.14,
       "step": 3000
     },
     {
-      "epoch": 1.54,
-      "learning_rate": 3.79401833092137e-05,
-      "loss": 2.7159,
+      "epoch": 3.08,
+      "learning_rate": 6.912055335968379e-05,
+      "loss": 2.0825,
       "step": 3500
     },
     {
-      "epoch": 1.54,
-      "eval_loss": 2.6153135299682617,
-      "eval_runtime": 81.5642,
-      "eval_samples_per_second": 191.052,
-      "eval_steps_per_second": 11.942,
+      "epoch": 3.08,
+      "eval_loss": 2.0795915126800537,
+      "eval_runtime": 79.1878,
+      "eval_samples_per_second": 196.785,
+      "eval_steps_per_second": 6.15,
       "step": 3500
     },
     {
-      "epoch": 1.76,
-      "learning_rate": 3.552821997105644e-05,
-      "loss": 2.6086,
+      "epoch": 3.52,
+      "learning_rate": 6.294466403162056e-05,
+      "loss": 1.9298,
       "step": 4000
     },
     {
-      "epoch": 1.76,
-      "eval_loss": 2.522780418395996,
-      "eval_runtime": 81.5345,
-      "eval_samples_per_second": 191.122,
-      "eval_steps_per_second": 11.946,
+      "epoch": 3.52,
+      "eval_loss": 2.0210442543029785,
+      "eval_runtime": 79.2843,
+      "eval_samples_per_second": 196.546,
+      "eval_steps_per_second": 6.142,
       "step": 4000
     },
     {
-      "epoch": 1.98,
-      "learning_rate": 3.311625663289918e-05,
-      "loss": 2.52,
+      "epoch": 3.96,
+      "learning_rate": 5.6768774703557316e-05,
+      "loss": 1.8936,
       "step": 4500
     },
     {
-      "epoch": 1.98,
-      "eval_loss": 2.446049451828003,
-      "eval_runtime": 81.2437,
-      "eval_samples_per_second": 191.806,
-      "eval_steps_per_second": 11.989,
+      "epoch": 3.96,
+      "eval_loss": 1.966531753540039,
+      "eval_runtime": 79.2784,
+      "eval_samples_per_second": 196.561,
+      "eval_steps_per_second": 6.143,
       "step": 4500
     },
     {
-      "epoch": 2.2,
-      "learning_rate": 3.0704293294741923e-05,
-      "loss": 2.3734,
+      "epoch": 4.4,
+      "learning_rate": 5.059288537549407e-05,
+      "loss": 1.746,
       "step": 5000
     },
     {
-      "epoch": 2.2,
-      "eval_loss": 2.3772902488708496,
-      "eval_runtime": 81.4154,
-      "eval_samples_per_second": 191.401,
-      "eval_steps_per_second": 11.963,
+      "epoch": 4.4,
+      "eval_loss": 1.9314507246017456,
+      "eval_runtime": 79.2451,
+      "eval_samples_per_second": 196.643,
+      "eval_steps_per_second": 6.145,
       "step": 5000
     },
     {
-      "epoch": 2.42,
-      "learning_rate": 2.829232995658466e-05,
-      "loss": 2.3257,
+      "epoch": 4.84,
+      "learning_rate": 4.441699604743083e-05,
+      "loss": 1.717,
       "step": 5500
     },
     {
-      "epoch": 2.42,
-      "eval_loss": 2.322295904159546,
-      "eval_runtime": 81.2639,
-      "eval_samples_per_second": 191.758,
-      "eval_steps_per_second": 11.986,
+      "epoch": 4.84,
+      "eval_loss": 1.8987737894058228,
+      "eval_runtime": 79.2498,
+      "eval_samples_per_second": 196.631,
+      "eval_steps_per_second": 6.145,
       "step": 5500
     },
     {
-      "epoch": 2.64,
-      "learning_rate": 2.58803666184274e-05,
-      "loss": 2.2788,
+      "epoch": 5.28,
+      "learning_rate": 3.824110671936759e-05,
+      "loss": 1.6156,
       "step": 6000
     },
     {
-      "epoch": 2.64,
-      "eval_loss": 2.274811267852783,
-      "eval_runtime": 81.4351,
-      "eval_samples_per_second": 191.355,
-      "eval_steps_per_second": 11.96,
+      "epoch": 5.28,
+      "eval_loss": 1.8819612264633179,
+      "eval_runtime": 79.3489,
+      "eval_samples_per_second": 196.386,
+      "eval_steps_per_second": 6.137,
       "step": 6000
     },
     {
-      "epoch": 2.86,
-      "learning_rate": 2.346840328027014e-05,
-      "loss": 2.2449,
+      "epoch": 5.72,
+      "learning_rate": 3.2065217391304345e-05,
+      "loss": 1.5674,
       "step": 6500
     },
     {
-      "epoch": 2.86,
-      "eval_loss": 2.233612298965454,
-      "eval_runtime": 81.4361,
-      "eval_samples_per_second": 191.352,
-      "eval_steps_per_second": 11.96,
+      "epoch": 5.72,
+      "eval_loss": 1.8578243255615234,
+      "eval_runtime": 79.2896,
+      "eval_samples_per_second": 196.533,
+      "eval_steps_per_second": 6.142,
       "step": 6500
     },
     {
-      "epoch": 3.08,
-      "learning_rate": 2.105643994211288e-05,
-      "loss": 2.1734,
+      "epoch": 6.16,
+      "learning_rate": 2.588932806324111e-05,
+      "loss": 1.5229,
       "step": 7000
     },
     {
-      "epoch": 3.08,
-      "eval_loss": 2.195352077484131,
-      "eval_runtime": 81.4409,
-      "eval_samples_per_second": 191.341,
-      "eval_steps_per_second": 11.96,
+      "epoch": 6.16,
+      "eval_loss": 1.848434567451477,
+      "eval_runtime": 79.2658,
+      "eval_samples_per_second": 196.592,
+      "eval_steps_per_second": 6.144,
       "step": 7000
     },
     {
-      "epoch": 3.3,
-      "learning_rate": 1.864447660395562e-05,
-      "loss": 2.088,
+      "epoch": 6.6,
+      "learning_rate": 1.9713438735177866e-05,
+      "loss": 1.4411,
       "step": 7500
     },
     {
-      "epoch": 3.3,
-      "eval_loss": 2.172189950942993,
-      "eval_runtime": 81.2625,
-      "eval_samples_per_second": 191.761,
-      "eval_steps_per_second": 11.986,
+      "epoch": 6.6,
+      "eval_loss": 1.8373626470565796,
+      "eval_runtime": 79.267,
+      "eval_samples_per_second": 196.589,
+      "eval_steps_per_second": 6.144,
       "step": 7500
     },
     {
-      "epoch": 3.52,
-      "learning_rate": 1.623251326579836e-05,
-      "loss": 2.0633,
+      "epoch": 7.04,
+      "learning_rate": 1.3537549407114625e-05,
+      "loss": 1.4288,
       "step": 8000
     },
     {
-      "epoch": 3.52,
-      "eval_loss": 2.1411869525909424,
-      "eval_runtime": 81.2032,
-      "eval_samples_per_second": 191.901,
-      "eval_steps_per_second": 11.995,
+      "epoch": 7.04,
+      "eval_loss": 1.8292683362960815,
+      "eval_runtime": 79.3154,
+      "eval_samples_per_second": 196.469,
+      "eval_steps_per_second": 6.14,
       "step": 8000
     },
     {
-      "epoch": 3.74,
-      "learning_rate": 1.3825373854317415e-05,
-      "loss": 2.0491,
+      "epoch": 7.48,
+      "learning_rate": 7.361660079051384e-06,
+      "loss": 1.3433,
       "step": 8500
     },
     {
-      "epoch": 3.74,
-      "eval_loss": 2.117720127105713,
-      "eval_runtime": 81.4728,
-      "eval_samples_per_second": 191.266,
-      "eval_steps_per_second": 11.955,
+      "epoch": 7.48,
+      "eval_loss": 1.8282458782196045,
+      "eval_runtime": 79.2978,
+      "eval_samples_per_second": 196.512,
+      "eval_steps_per_second": 6.141,
       "step": 8500
     },
     {
-      "epoch": 3.96,
-      "learning_rate": 1.1413410516160154e-05,
-      "loss": 2.032,
+      "epoch": 7.92,
+      "learning_rate": 1.1857707509881422e-06,
+      "loss": 1.3415,
       "step": 9000
     },
     {
-      "epoch": 3.96,
-      "eval_loss": 2.0970561504364014,
-      "eval_runtime": 81.4071,
-      "eval_samples_per_second": 191.421,
-      "eval_steps_per_second": 11.965,
+      "epoch": 7.92,
+      "eval_loss": 1.8225059509277344,
+      "eval_runtime": 79.2921,
+      "eval_samples_per_second": 196.527,
+      "eval_steps_per_second": 6.142,
       "step": 9000
-    },
-    {
-      "epoch": 4.18,
-      "learning_rate": 9.001447178002895e-06,
-      "loss": 1.9451,
-      "step": 9500
-    },
-    {
-      "epoch": 4.18,
-      "eval_loss": 2.084089756011963,
-      "eval_runtime": 81.4657,
-      "eval_samples_per_second": 191.283,
-      "eval_steps_per_second": 11.956,
-      "step": 9500
-    },
-    {
-      "epoch": 4.4,
-      "learning_rate": 6.589483839845635e-06,
-      "loss": 1.9275,
-      "step": 10000
-    },
-    {
-      "epoch": 4.4,
-      "eval_loss": 2.0687994956970215,
-      "eval_runtime": 81.6182,
-      "eval_samples_per_second": 190.926,
-      "eval_steps_per_second": 11.934,
-      "step": 10000
-    },
-    {
-      "epoch": 4.62,
-      "learning_rate": 4.18234442836469e-06,
-      "loss": 1.9138,
-      "step": 10500
-    },
-    {
-      "epoch": 4.62,
-      "eval_loss": 2.0576138496398926,
-      "eval_runtime": 81.6062,
-      "eval_samples_per_second": 190.954,
-      "eval_steps_per_second": 11.935,
-      "step": 10500
-    },
-    {
-      "epoch": 4.84,
-      "learning_rate": 1.7703810902074289e-06,
-      "loss": 1.9068,
-      "step": 11000
-    },
-    {
-      "epoch": 4.84,
-      "eval_loss": 2.048520565032959,
-      "eval_runtime": 81.6161,
-      "eval_samples_per_second": 190.93,
-      "eval_steps_per_second": 11.934,
-      "step": 11000
     }
   ],
-  "max_steps": 11365,
-  "num_train_epochs": 5,
-  "total_flos": 1.1494473283584e+16,
+  "max_steps": 9096,
+  "num_train_epochs": 8,
+  "total_flos": 1.88015947776e+16,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ff4a22b53672fbc7f2d5cc3310f4e17058a7a012b48b927465d691e9f95a218b
+oid sha256:9a43384642115ae1e433db04ff31d33786becdd7d0ea11e8c69f317f3f149bbd
 size 3963