Training in progress, step 2000, checkpoint

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:481549fd92bfa33b3ca3ee59d5bc352b3a2c5f4809eb649581a381291df624cc
+oid sha256:3e844c140f0b364ef240d9d283f3474471aa887e9a4e7f8afe6ca0d5062fa1d8
 size 847599616

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c4a7cf0ed2c81fc32c151e4b208e492f66b21183d21c3703bc7d39409f809f7a
+oid sha256:d6e57c8be8d618705f188b94fc5169164f579e23fc7d687c76672146f866dfc5
 size 1386414411

last-checkpoint/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cd45c472c02f69686aade1ec7cee7512c8dba6e6bde5e0bd023fd4f8f2f3548
+size 14917

last-checkpoint/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09714a2ef70dd987c21ebd404103aea88be622e980ab90fe03c4cb22a4d6d0de
+size 14917

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0008eaf86467511a3b74b3792388bc89587db0af5fba1e98578d98def9815f59
+oid sha256:c0ed1549b52af39b27c28b24276ca81b4d37dad6af883ad09a3c1268b5bcb1b0
 size 1465

last-checkpoint/trainer_state.json

@@ -2,210 +2,174 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.01,
-  "eval_steps": 300,
+  "epoch": 0.13333333333333333,
+  "eval_steps": 1000,
   "global_step": 2000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0005,
-      "grad_norm": 6.011565208435059,
-      "learning_rate": 3.96e-06,
-      "loss": 20.981669921875,
+      "epoch": 0.006666666666666667,
+      "grad_norm": 2.854929208755493,
+      "learning_rate": 5.28e-05,
+      "loss": 18.5724072265625,
       "step": 100
     },
     {
-      "epoch": 0.001,
-      "grad_norm": 4.79960823059082,
-      "learning_rate": 7.96e-06,
-      "loss": 18.630086669921877,
+      "epoch": 0.013333333333333334,
+      "grad_norm": 1.6583245992660522,
+      "learning_rate": 0.00010613333333333333,
+      "loss": 14.702509765625,
       "step": 200
     },
     {
-      "epoch": 0.0015,
-      "grad_norm": 4.753690242767334,
-      "learning_rate": 1.196e-05,
-      "loss": 17.497738037109375,
+      "epoch": 0.02,
+      "grad_norm": 1.6954909563064575,
+      "learning_rate": 0.00015946666666666668,
+      "loss": 13.69074951171875,
       "step": 300
     },
     {
-      "epoch": 0.0015,
-      "eval_accuracy": 0.08108023483365949,
-      "eval_loss": 16.94498062133789,
-      "eval_runtime": 72.1019,
-      "eval_samples_per_second": 13.869,
-      "eval_steps_per_second": 3.467,
-      "step": 300
-    },
-    {
-      "epoch": 0.002,
-      "grad_norm": 3.2824501991271973,
-      "learning_rate": 1.596e-05,
-      "loss": 16.3246142578125,
+      "epoch": 0.02666666666666667,
+      "grad_norm": 1.783247470855713,
+      "learning_rate": 0.00021280000000000002,
+      "loss": 13.1521044921875,
       "step": 400
     },
     {
-      "epoch": 0.0025,
-      "grad_norm": 4.837945461273193,
-      "learning_rate": 1.9960000000000002e-05,
-      "loss": 15.711016845703124,
+      "epoch": 0.03333333333333333,
+      "grad_norm": 1.4421945810317993,
+      "learning_rate": 0.00026613333333333337,
+      "loss": 12.824429931640625,
       "step": 500
     },
     {
-      "epoch": 0.003,
-      "grad_norm": 4.028888702392578,
-      "learning_rate": 2.396e-05,
-      "loss": 14.91444580078125,
+      "epoch": 0.04,
+      "grad_norm": 1.5354344844818115,
+      "learning_rate": 0.00031946666666666666,
+      "loss": 12.49961181640625,
       "step": 600
     },
     {
-      "epoch": 0.003,
-      "eval_accuracy": 0.10803913894324853,
-      "eval_loss": 14.84846305847168,
-      "eval_runtime": 71.0639,
-      "eval_samples_per_second": 14.072,
-      "eval_steps_per_second": 3.518,
-      "step": 600
-    },
-    {
-      "epoch": 0.0035,
-      "grad_norm": 2.6233296394348145,
-      "learning_rate": 2.7960000000000003e-05,
-      "loss": 14.73691650390625,
+      "epoch": 0.04666666666666667,
+      "grad_norm": 1.3272199630737305,
+      "learning_rate": 0.00037280000000000006,
+      "loss": 12.145262451171876,
       "step": 700
     },
     {
-      "epoch": 0.004,
-      "grad_norm": 3.3784563541412354,
-      "learning_rate": 3.196e-05,
-      "loss": 14.402127685546875,
+      "epoch": 0.05333333333333334,
+      "grad_norm": 1.706987977027893,
+      "learning_rate": 0.0003999883303467061,
+      "loss": 11.98048095703125,
       "step": 800
     },
     {
-      "epoch": 0.0045,
-      "grad_norm": 4.078140735626221,
-      "learning_rate": 3.596e-05,
-      "loss": 14.276339111328125,
+      "epoch": 0.06,
+      "grad_norm": 1.2479976415634155,
+      "learning_rate": 0.00039989210445800615,
+      "loss": 11.621953125,
       "step": 900
     },
     {
-      "epoch": 0.0045,
-      "eval_accuracy": 0.12540117416829746,
-      "eval_loss": 14.246834754943848,
-      "eval_runtime": 70.5499,
-      "eval_samples_per_second": 14.174,
-      "eval_steps_per_second": 3.544,
-      "step": 900
+      "epoch": 0.06666666666666667,
+      "grad_norm": 1.2310028076171875,
+      "learning_rate": 0.00039969872739228826,
+      "loss": 11.573458251953125,
+      "step": 1000
     },
     {
-      "epoch": 0.005,
-      "grad_norm": 2.698321580886841,
-      "learning_rate": 3.9960000000000004e-05,
-      "loss": 14.103670654296875,
+      "epoch": 0.06666666666666667,
+      "eval_accuracy": 0.18273972602739727,
+      "eval_loss": 11.462567329406738,
+      "eval_runtime": 65.4243,
+      "eval_samples_per_second": 7.642,
+      "eval_steps_per_second": 0.382,
       "step": 1000
     },
     {
-      "epoch": 0.0055,
-      "grad_norm": 2.6977009773254395,
-      "learning_rate": 4.396e-05,
-      "loss": 13.91288330078125,
+      "epoch": 0.07333333333333333,
+      "grad_norm": 1.2715719938278198,
+      "learning_rate": 0.00039940829313430293,
+      "loss": 11.32210205078125,
       "step": 1100
     },
     {
-      "epoch": 0.006,
-      "grad_norm": 2.813852071762085,
-      "learning_rate": 4.796e-05,
-      "loss": 13.750213623046875,
-      "step": 1200
-    },
-    {
-      "epoch": 0.006,
-      "eval_accuracy": 0.13354794520547944,
-      "eval_loss": 13.796219825744629,
-      "eval_runtime": 70.4436,
-      "eval_samples_per_second": 14.196,
-      "eval_steps_per_second": 3.549,
+      "epoch": 0.08,
+      "grad_norm": 1.7137954235076904,
+      "learning_rate": 0.00039902094284035075,
+      "loss": 11.18190185546875,
       "step": 1200
     },
     {
-      "epoch": 0.0065,
-      "grad_norm": 2.7015318870544434,
-      "learning_rate": 5.196e-05,
-      "loss": 13.593814697265625,
+      "epoch": 0.08666666666666667,
+      "grad_norm": 1.48090660572052,
+      "learning_rate": 0.0003985368647696784,
+      "loss": 11.0220166015625,
       "step": 1300
     },
     {
-      "epoch": 0.007,
-      "grad_norm": 2.6103267669677734,
-      "learning_rate": 5.596e-05,
-      "loss": 13.418026123046875,
+      "epoch": 0.09333333333333334,
+      "grad_norm": 1.1374964714050293,
+      "learning_rate": 0.0003979562941929808,
+      "loss": 10.822305908203125,
       "step": 1400
     },
     {
-      "epoch": 0.0075,
-      "grad_norm": 2.5891685485839844,
-      "learning_rate": 5.996e-05,
-      "loss": 13.413653564453124,
-      "step": 1500
-    },
-    {
-      "epoch": 0.0075,
-      "eval_accuracy": 0.14203326810176126,
-      "eval_loss": 13.424027442932129,
-      "eval_runtime": 70.7197,
-      "eval_samples_per_second": 14.14,
-      "eval_steps_per_second": 3.535,
+      "epoch": 0.1,
+      "grad_norm": 1.3714447021484375,
+      "learning_rate": 0.0003972795132780554,
+      "loss": 10.762421875,
       "step": 1500
     },
     {
-      "epoch": 0.008,
-      "grad_norm": 2.7721621990203857,
-      "learning_rate": 6.396e-05,
-      "loss": 13.18080810546875,
+      "epoch": 0.10666666666666667,
+      "grad_norm": 1.1782511472702026,
+      "learning_rate": 0.00039650685095266405,
+      "loss": 10.602275390625,
       "step": 1600
     },
     {
-      "epoch": 0.0085,
-      "grad_norm": 4.532048225402832,
-      "learning_rate": 6.796e-05,
-      "loss": 13.176103515625,
+      "epoch": 0.11333333333333333,
+      "grad_norm": 1.0709394216537476,
+      "learning_rate": 0.0003956386827446671,
+      "loss": 10.4713134765625,
       "step": 1700
     },
     {
-      "epoch": 0.009,
-      "grad_norm": 2.760545015335083,
-      "learning_rate": 7.196000000000001e-05,
-      "loss": 13.289417724609375,
+      "epoch": 0.12,
+      "grad_norm": 1.1138310432434082,
+      "learning_rate": 0.00039467543059951106,
+      "loss": 10.50548095703125,
       "step": 1800
     },
     {
-      "epoch": 0.009,
-      "eval_accuracy": 0.1470078277886497,
-      "eval_loss": 13.157907485961914,
-      "eval_runtime": 70.4071,
-      "eval_samples_per_second": 14.203,
-      "eval_steps_per_second": 3.551,
-      "step": 1800
+      "epoch": 0.12666666666666668,
+      "grad_norm": 1.283762812614441,
+      "learning_rate": 0.0003936175626751552,
+      "loss": 10.4138134765625,
+      "step": 1900
     },
     {
-      "epoch": 0.0095,
-      "grad_norm": 2.832441806793213,
-      "learning_rate": 7.596000000000001e-05,
-      "loss": 12.94510009765625,
-      "step": 1900
+      "epoch": 0.13333333333333333,
+      "grad_norm": 1.3275446891784668,
+      "learning_rate": 0.00039246559311453794,
+      "loss": 10.335478515625,
+      "step": 2000
     },
     {
-      "epoch": 0.01,
-      "grad_norm": 2.487884759902954,
-      "learning_rate": 7.996e-05,
-      "loss": 12.798345947265625,
+      "epoch": 0.13333333333333333,
+      "eval_accuracy": 0.21337964774951076,
+      "eval_loss": 10.411324501037598,
+      "eval_runtime": 65.3972,
+      "eval_samples_per_second": 7.646,
+      "eval_steps_per_second": 0.382,
       "step": 2000
     }
   ],
   "logging_steps": 100,
-  "max_steps": 200000,
+  "max_steps": 15000,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 9223372036854775807,
   "save_steps": 2000,
@@ -222,7 +186,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 4,
+  "train_batch_size": 10,
   "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:d28e2b5843184725d1b8a4fbebe4630c5d5705edaadbc1cb4642d87884e1e3e9
+oid sha256:f471a8fab1d3fe9130aacb1db6fd8b7b3e66edecb9210651c26dd497a9c1abea
 size 5265