Training in progress, step 20000

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a9b46ed6af836e8924481b9ef3b8ac908f671a8de3334b85575e607c66e5a72
+oid sha256:65debe850847edf2f62bba0deb29483af297ffa34c6cdcd66a55832044fcc8ec
 size 202193937

last-checkpoint/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9fa1850686699db6f7d38adef4e1f7555c082f72c52d9543757d96d6e7dfe575
+oid sha256:e6f55f7fb7116bd42203f7b9313d6b9f08f6a023278c3f67be7ce3fba873b5dc
 size 102501541

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/rng_state_4.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/rng_state_5.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/rng_state_6.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/rng_state_7.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f63b6e7591877af348d0715da8e499a40da21a9ac10ea61d3d17f2882b7374e
+oid sha256:006a29e4d6e5dacaa8f0d3ba56263cd7f1775a1129713c79ccc62a2c04246894
 size 14503

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4931c97be633a6e3808391ac03b792837805ded64ff2a578babfa79e60dc22a7
+oid sha256:c84272d9e8a15cd78c706d16b77833d53dbfbd182e8ad79e3cd658ef6c3eaaf6
 size 623

last-checkpoint/trainer_state.json

@@ -1,8 +1,8 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.23299704093758009,
-  "global_step": 10000,
+  "epoch": 0.46599408187516017,
+  "global_step": 20000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -206,11 +206,211 @@
       "eval_samples_per_second": 1737.882,
       "eval_steps_per_second": 27.673,
       "step": 10000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00012599999999999997,
+      "loss": 0.5516,
+      "step": 10500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00013199999999999998,
+      "loss": 0.5428,
+      "step": 11000
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 0.5300613045692444,
+      "eval_runtime": 1.2658,
+      "eval_samples_per_second": 1736.485,
+      "eval_steps_per_second": 27.651,
+      "step": 11000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.000138,
+      "loss": 0.5346,
+      "step": 11500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00014399999999999998,
+      "loss": 0.527,
+      "step": 12000
+    },
+    {
+      "epoch": 0.28,
+      "eval_loss": 0.5140743255615234,
+      "eval_runtime": 1.2609,
+      "eval_samples_per_second": 1743.192,
+      "eval_steps_per_second": 27.758,
+      "step": 12000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00015,
+      "loss": 0.5203,
+      "step": 12500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.000156,
+      "loss": 0.5142,
+      "step": 13000
+    },
+    {
+      "epoch": 0.3,
+      "eval_loss": 0.5022566318511963,
+      "eval_runtime": 1.2481,
+      "eval_samples_per_second": 1761.066,
+      "eval_steps_per_second": 28.042,
+      "step": 13000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.000162,
+      "loss": 0.5079,
+      "step": 13500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.000168,
+      "loss": 0.5002,
+      "step": 14000
+    },
+    {
+      "epoch": 0.33,
+      "eval_loss": 0.4841987192630768,
+      "eval_runtime": 1.3176,
+      "eval_samples_per_second": 1668.178,
+      "eval_steps_per_second": 26.563,
+      "step": 14000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00017399999999999997,
+      "loss": 0.4916,
+      "step": 14500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00017999999999999998,
+      "loss": 0.4852,
+      "step": 15000
+    },
+    {
+      "epoch": 0.35,
+      "eval_loss": 0.47066375613212585,
+      "eval_runtime": 1.3026,
+      "eval_samples_per_second": 1687.422,
+      "eval_steps_per_second": 26.87,
+      "step": 15000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.000186,
+      "loss": 0.4778,
+      "step": 15500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00019199999999999998,
+      "loss": 0.4697,
+      "step": 16000
+    },
+    {
+      "epoch": 0.37,
+      "eval_loss": 0.4562186300754547,
+      "eval_runtime": 1.2895,
+      "eval_samples_per_second": 1704.503,
+      "eval_steps_per_second": 27.142,
+      "step": 16000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.000198,
+      "loss": 0.4636,
+      "step": 16500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.000204,
+      "loss": 0.459,
+      "step": 17000
+    },
+    {
+      "epoch": 0.4,
+      "eval_loss": 0.44295796751976013,
+      "eval_runtime": 1.2516,
+      "eval_samples_per_second": 1756.093,
+      "eval_steps_per_second": 27.963,
+      "step": 17000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00020999999999999998,
+      "loss": 0.4544,
+      "step": 17500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00021599999999999996,
+      "loss": 0.45,
+      "step": 18000
+    },
+    {
+      "epoch": 0.42,
+      "eval_loss": 0.4343169629573822,
+      "eval_runtime": 1.327,
+      "eval_samples_per_second": 1656.351,
+      "eval_steps_per_second": 26.375,
+      "step": 18000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00022199999999999998,
+      "loss": 0.4458,
+      "step": 18500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00022799999999999999,
+      "loss": 0.4411,
+      "step": 19000
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 0.42496559023857117,
+      "eval_runtime": 1.3285,
+      "eval_samples_per_second": 1654.492,
+      "eval_steps_per_second": 26.345,
+      "step": 19000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.000234,
+      "loss": 0.4368,
+      "step": 19500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00023999999999999998,
+      "loss": 0.4322,
+      "step": 20000
+    },
+    {
+      "epoch": 0.47,
+      "eval_loss": 0.4132905900478363,
+      "eval_runtime": 1.3158,
+      "eval_samples_per_second": 1670.425,
+      "eval_steps_per_second": 26.599,
+      "step": 20000
     }
   ],
   "max_steps": 500000,
   "num_train_epochs": 12,
-  "total_flos": 3.194871387745e+20,
+  "total_flos": 6.38974277549e+20,
   "trial_name": null,
   "trial_params": null
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9fa1850686699db6f7d38adef4e1f7555c082f72c52d9543757d96d6e7dfe575
+oid sha256:e6f55f7fb7116bd42203f7b9313d6b9f08f6a023278c3f67be7ce3fba873b5dc
 size 102501541