Training in progress, step 4000, checkpoint

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1069d6130a8bebfd2ceeda1d0936f626f5637e1fc5ef945629ede330c4c05e04
-size 812034507
+oid sha256:fe373abbfe5cca2334d2427e9171b4bb2c3f4a319ffc60e24e707b0beb7e2fc4
+size 1442699

last-checkpoint/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a3234ebdb3c9bd5374463f6ff57f5856cedb2a1db06c4a225719dc1fe2d4e844
+oid sha256:36693100ec9169174718c3c2f2fd50d9d486ea7cc9d06d26b75681fb6c2f97bd
 size 499774435

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:799c2a9b233aca9b1649a975d8fc2d128798bc9a26769fb0f6cf72a67c9a46d2
+oid sha256:851f1a1d183e5897767447c672bcf8362d9ae2a4736cbb50b4e208139d171c19
 size 14469

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:89e9f7fe1ab29af6aa77e6004c57876d5313c9d3da26e9e11a0acfa75ab42122
+oid sha256:a7fa122671d1ef4d22de37e01b14987409ecd975ea71c5bf79191d6ce98cf8a3
 size 14469

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7fff2ccf7c72f15d3b76ddb019c26a4dca9d2e1f2ee32840db12f9fec2b95bea
+oid sha256:0836353fc5b4695271900d529f17e5c9bcc8f6b960c6c8ad2842449359842642
 size 14469

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:83a4e867fa1241d92d6bcf27766baa9f317570cbfb999ddb76e020d43ec7d4dc
+oid sha256:d85458cac81ee63dbddcc21880d9d5d65d2999a521682ced8bb92cd38533d72b
 size 14469

last-checkpoint/rng_state_4.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f4acf2b09a1339e370773a0fd0c75b1d5891aee4c56e1dd818e17f835520caf
+oid sha256:5b61eff2b0a5194e1f7cf6b678cdf994e3194534b171d100afdd85158f083dbf
 size 14469

last-checkpoint/rng_state_5.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7546ae9c1e6fa6e6426b888392672f7514267f83e14e93ef8fc7db13a5220a3b
+oid sha256:9e09f720647dea58d44bd8a7064bc267a53d19c77ef833401671fb959ec0bee3
 size 14469

last-checkpoint/rng_state_6.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f84e63e29d923cc8c21554fe213de1a06f00b47198b99fe493dcf59f2ace055
+oid sha256:4fc874467b18473948fa002d73f9ced910bc93328612c168e89eefd3ff25af66
 size 14469

last-checkpoint/rng_state_7.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7b48f05a6fe6d1fbe948bd035323503b6cde575c8bc475d73338ea1f2101f32e
+oid sha256:e642abe2620703f2d27e2aefa3f96715c8ae63a82165f9e63e498f488c2c681d
 size 14469

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eb2a89a85e1c8ac7ee9b5678d857a4d99090faa37a24489f44839e0801dfe16a
+oid sha256:76cf98a8cde877f1e49a9e4b884c4dce3d51bb4035429d464e74acc35ddca024
 size 1465

last-checkpoint/trainer_state.json

@@ -2,486 +2,328 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.12,
+  "epoch": 0.08,
   "eval_steps": 1000,
-  "global_step": 6000,
+  "global_step": 4000,
   "is_hyper_param_search": false,
-  "is_local_process_zero": true,
+  "is_local_process_zero": false,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.002,
-      "grad_norm": 1.2400550842285156,
+      "grad_norm": 19.660429000854492,
       "learning_rate": 7.920000000000001e-05,
-      "loss": 8.362694702148438,
+      "loss": 67.8543,
       "step": 100
     },
     {
       "epoch": 0.004,
-      "grad_norm": 2.1569149494171143,
+      "grad_norm": 24.802082061767578,
       "learning_rate": 0.00015920000000000002,
-      "loss": 6.592593383789063,
+      "loss": 54.1444,
       "step": 200
     },
     {
       "epoch": 0.006,
-      "grad_norm": 4.070979595184326,
+      "grad_norm": 16.059066772460938,
       "learning_rate": 0.00023920000000000001,
-      "loss": 6.208985595703125,
+      "loss": 51.5921,
       "step": 300
     },
     {
       "epoch": 0.008,
-      "grad_norm": 4.392622470855713,
+      "grad_norm": 27.705520629882812,
       "learning_rate": 0.0003192,
-      "loss": 5.921519775390625,
+      "loss": 50.0536,
       "step": 400
     },
     {
       "epoch": 0.01,
-      "grad_norm": 4.653286457061768,
+      "grad_norm": 25.79052734375,
       "learning_rate": 0.0003992,
-      "loss": 5.658856201171875,
+      "loss": 48.6987,
       "step": 500
     },
     {
       "epoch": 0.012,
-      "grad_norm": 2.6610119342803955,
+      "grad_norm": 14.82126522064209,
       "learning_rate": 0.00047920000000000005,
-      "loss": 5.450134887695312,
+      "loss": 47.3182,
       "step": 600
     },
     {
       "epoch": 0.014,
-      "grad_norm": 2.6298203468322754,
+      "grad_norm": 12.00988483428955,
       "learning_rate": 0.0005592,
-      "loss": 5.260469360351562,
+      "loss": 46.6574,
       "step": 700
     },
     {
       "epoch": 0.016,
-      "grad_norm": 1.681970238685608,
+      "grad_norm": 32.48664474487305,
       "learning_rate": 0.0006392,
-      "loss": 5.112535095214843,
+      "loss": 46.0747,
       "step": 800
     },
     {
       "epoch": 0.018,
-      "grad_norm": 1.5673184394836426,
+      "grad_norm": 25.646282196044922,
       "learning_rate": 0.0007191999999999999,
-      "loss": 5.0076907348632815,
+      "loss": 45.6098,
       "step": 900
     },
     {
       "epoch": 0.02,
-      "grad_norm": 1.6654455661773682,
+      "grad_norm": 16.071203231811523,
       "learning_rate": 0.0007992,
-      "loss": 4.9159228515625,
+      "loss": 44.8537,
       "step": 1000
     },
     {
       "epoch": 0.02,
-      "eval_accuracy": 0.2274637964774951,
-      "eval_loss": 4.8702616691589355,
-      "eval_runtime": 5.1375,
-      "eval_samples_per_second": 194.646,
-      "eval_steps_per_second": 1.557,
+      "eval_accuracy": 0.18606457925636008,
+      "eval_loss": 44.18140411376953,
+      "eval_runtime": 14.581,
+      "eval_samples_per_second": 8.573,
+      "eval_steps_per_second": 0.137,
       "step": 1000
     },
     {
       "epoch": 0.022,
-      "grad_norm": 1.3562606573104858,
+      "grad_norm": 14.277446746826172,
       "learning_rate": 0.0008792,
-      "loss": 4.811459045410157,
+      "loss": 44.223,
       "step": 1100
     },
     {
       "epoch": 0.024,
-      "grad_norm": 1.1230770349502563,
+      "grad_norm": 9.809870719909668,
       "learning_rate": 0.0009592000000000001,
-      "loss": 4.759424438476563,
+      "loss": 43.5719,
       "step": 1200
     },
     {
       "epoch": 0.026,
-      "grad_norm": 0.9026045203208923,
+      "grad_norm": 16.994014739990234,
       "learning_rate": 0.0010391999999999999,
-      "loss": 4.691336669921875,
+      "loss": 43.0323,
       "step": 1300
     },
     {
       "epoch": 0.028,
-      "grad_norm": 0.8914104104042053,
+      "grad_norm": 8.082005500793457,
       "learning_rate": 0.0011192,
-      "loss": 4.6415283203125,
+      "loss": 42.3631,
       "step": 1400
     },
     {
       "epoch": 0.03,
-      "grad_norm": 0.8752658367156982,
+      "grad_norm": 7.080636024475098,
       "learning_rate": 0.0011992,
-      "loss": 4.560531311035156,
+      "loss": 42.0285,
       "step": 1500
     },
     {
       "epoch": 0.032,
-      "grad_norm": 0.7435582280158997,
+      "grad_norm": 8.547815322875977,
       "learning_rate": 0.0012791999999999999,
-      "loss": 4.516304016113281,
+      "loss": 42.6829,
       "step": 1600
     },
     {
       "epoch": 0.034,
-      "grad_norm": 0.7041499018669128,
+      "grad_norm": 7.541939735412598,
       "learning_rate": 0.0013592,
-      "loss": 4.48115234375,
+      "loss": 41.9796,
       "step": 1700
     },
     {
       "epoch": 0.036,
-      "grad_norm": 0.5327283143997192,
+      "grad_norm": 6.902923583984375,
       "learning_rate": 0.0014392,
-      "loss": 4.429443969726562,
+      "loss": 41.5016,
       "step": 1800
     },
     {
       "epoch": 0.038,
-      "grad_norm": 0.45207348465919495,
+      "grad_norm": 7.68479061126709,
       "learning_rate": 0.0015192,
-      "loss": 4.37362060546875,
+      "loss": 41.2607,
       "step": 1900
     },
     {
       "epoch": 0.04,
-      "grad_norm": 0.5152547359466553,
+      "grad_norm": 8.936702728271484,
       "learning_rate": 0.0015992,
-      "loss": 4.377236938476562,
+      "loss": 40.7545,
       "step": 2000
     },
     {
       "epoch": 0.04,
-      "eval_accuracy": 0.2694227005870842,
-      "eval_loss": 4.3582682609558105,
-      "eval_runtime": 5.3878,
-      "eval_samples_per_second": 185.606,
-      "eval_steps_per_second": 1.485,
+      "eval_accuracy": 0.21657142857142858,
+      "eval_loss": 40.417686462402344,
+      "eval_runtime": 3.051,
+      "eval_samples_per_second": 40.97,
+      "eval_steps_per_second": 0.656,
       "step": 2000
     },
     {
       "epoch": 0.042,
-      "grad_norm": 0.45765218138694763,
+      "grad_norm": 5.834764003753662,
       "learning_rate": 0.0016792,
-      "loss": 4.341053161621094,
+      "loss": 40.2024,
       "step": 2100
     },
     {
       "epoch": 0.044,
-      "grad_norm": 0.3776390254497528,
+      "grad_norm": 3.9864397048950195,
       "learning_rate": 0.0017592,
-      "loss": 4.304255676269531,
+      "loss": 39.4534,
       "step": 2200
     },
     {
       "epoch": 0.046,
-      "grad_norm": 0.48194995522499084,
+      "grad_norm": 6.052990436553955,
       "learning_rate": 0.0018392,
-      "loss": 4.308780822753906,
+      "loss": 40.0088,
       "step": 2300
     },
     {
       "epoch": 0.048,
-      "grad_norm": 0.40603286027908325,
+      "grad_norm": 10.583518981933594,
       "learning_rate": 0.0019192,
-      "loss": 4.272806396484375,
+      "loss": 40.099,
       "step": 2400
     },
     {
       "epoch": 0.05,
-      "grad_norm": 0.41386568546295166,
+      "grad_norm": 3.686554193496704,
       "learning_rate": 0.0019992,
-      "loss": 4.267073974609375,
+      "loss": 39.5664,
       "step": 2500
     },
     {
       "epoch": 0.052,
-      "grad_norm": 0.4916844666004181,
+      "grad_norm": 2.981701612472534,
       "learning_rate": 0.001999978563623903,
-      "loss": 4.220490417480469,
+      "loss": 39.5087,
       "step": 2600
     },
     {
       "epoch": 0.054,
-      "grad_norm": 0.37477609515190125,
+      "grad_norm": 3.382185459136963,
       "learning_rate": 0.0019999133871331223,
-      "loss": 4.224794006347656,
+      "loss": 38.9185,
       "step": 2700
     },
     {
       "epoch": 0.056,
-      "grad_norm": 0.405086874961853,
+      "grad_norm": 2.8874025344848633,
       "learning_rate": 0.0019998044711915177,
-      "loss": 4.185610961914063,
+      "loss": 38.1472,
       "step": 2800
     },
     {
       "epoch": 0.058,
-      "grad_norm": 0.334581196308136,
+      "grad_norm": 3.083789825439453,
       "learning_rate": 0.0019996518205634257,
-      "loss": 4.165193176269531,
+      "loss": 37.5833,
       "step": 2900
     },
     {
       "epoch": 0.06,
-      "grad_norm": 0.3760662376880646,
+      "grad_norm": 2.315279483795166,
       "learning_rate": 0.0019994554419262797,
-      "loss": 4.161107177734375,
+      "loss": 37.927,
       "step": 3000
     },
     {
       "epoch": 0.06,
-      "eval_accuracy": 0.28686888454011744,
-      "eval_loss": 4.15764856338501,
-      "eval_runtime": 5.1036,
-      "eval_samples_per_second": 195.938,
-      "eval_steps_per_second": 1.568,
+      "eval_accuracy": 0.23828180039138944,
+      "eval_loss": 37.86294174194336,
+      "eval_runtime": 3.1437,
+      "eval_samples_per_second": 39.763,
+      "eval_steps_per_second": 0.636,
       "step": 3000
     },
     {
       "epoch": 0.062,
-      "grad_norm": 0.38909581303596497,
+      "grad_norm": 4.6366777420043945,
       "learning_rate": 0.001999215343870317,
-      "loss": 4.1489105224609375,
+      "loss": 38.432,
       "step": 3100
     },
     {
       "epoch": 0.064,
-      "grad_norm": 0.4108506143093109,
+      "grad_norm": 2.8195507526397705,
       "learning_rate": 0.0019989315368982054,
-      "loss": 4.123897705078125,
+      "loss": 37.8144,
       "step": 3200
     },
     {
       "epoch": 0.066,
-      "grad_norm": 0.27484413981437683,
+      "grad_norm": 2.320726156234741,
       "learning_rate": 0.0019986040334245797,
-      "loss": 4.101358032226562,
+      "loss": 38.1739,
       "step": 3300
     },
     {
       "epoch": 0.068,
-      "grad_norm": 0.29958581924438477,
+      "grad_norm": 2.575469732284546,
       "learning_rate": 0.001998232847775504,
-      "loss": 4.10423583984375,
+      "loss": 37.1996,
       "step": 3400
     },
     {
       "epoch": 0.07,
-      "grad_norm": 0.29483968019485474,
+      "grad_norm": 2.4692764282226562,
       "learning_rate": 0.0019978179961878404,
-      "loss": 4.078011169433593,
+      "loss": 36.8725,
       "step": 3500
     },
     {
       "epoch": 0.072,
-      "grad_norm": 0.36317598819732666,
+      "grad_norm": 2.555267810821533,
       "learning_rate": 0.001997359496808541,
-      "loss": 4.076428527832031,
+      "loss": 36.2038,
       "step": 3600
     },
     {
       "epoch": 0.074,
-      "grad_norm": 0.34998658299446106,
+      "grad_norm": 1.9989144802093506,
       "learning_rate": 0.001996857369693855,
-      "loss": 4.052371826171875,
+      "loss": 36.3164,
       "step": 3700
     },
     {
       "epoch": 0.076,
-      "grad_norm": 0.31070172786712646,
+      "grad_norm": 2.223381519317627,
       "learning_rate": 0.0019963116368084486,
-      "loss": 4.059148559570312,
+      "loss": 37.1086,
       "step": 3800
     },
     {
       "epoch": 0.078,
-      "grad_norm": 0.2655385732650757,
+      "grad_norm": 1.996813416481018,
       "learning_rate": 0.001995722322024446,
-      "loss": 4.039622802734375,
+      "loss": 36.5327,
       "step": 3900
     },
     {
       "epoch": 0.08,
-      "grad_norm": 0.4650065302848816,
+      "grad_norm": 1.8378113508224487,
       "learning_rate": 0.001995089451120385,
-      "loss": 4.011789855957031,
+      "loss": 36.314,
       "step": 4000
     },
     {
       "epoch": 0.08,
-      "eval_accuracy": 0.29829745596868884,
-      "eval_loss": 4.0341691970825195,
-      "eval_runtime": 5.1511,
-      "eval_samples_per_second": 194.133,
-      "eval_steps_per_second": 1.553,
+      "eval_accuracy": 0.2553561643835616,
+      "eval_loss": 36.08281707763672,
+      "eval_runtime": 3.1615,
+      "eval_samples_per_second": 39.538,
+      "eval_steps_per_second": 0.633,
       "step": 4000
-    },
-    {
-      "epoch": 0.082,
-      "grad_norm": 0.43114492297172546,
-      "learning_rate": 0.0019944130517800893,
-      "loss": 4.0238558959960935,
-      "step": 4100
-    },
-    {
-      "epoch": 0.084,
-      "grad_norm": 0.26026225090026855,
-      "learning_rate": 0.001993693153591457,
-      "loss": 4.017567443847656,
-      "step": 4200
-    },
-    {
-      "epoch": 0.086,
-      "grad_norm": 0.23712828755378723,
-      "learning_rate": 0.0019929297880451674,
-      "loss": 3.9972265625,
-      "step": 4300
-    },
-    {
-      "epoch": 0.088,
-      "grad_norm": 0.44415482878685,
-      "learning_rate": 0.0019921229885333023,
-      "loss": 3.9874578857421876,
-      "step": 4400
-    },
-    {
-      "epoch": 0.09,
-      "grad_norm": 0.27471208572387695,
-      "learning_rate": 0.001991272790347886,
-      "loss": 3.998799133300781,
-      "step": 4500
-    },
-    {
-      "epoch": 0.092,
-      "grad_norm": 0.2248983383178711,
-      "learning_rate": 0.0019903792306793415,
-      "loss": 3.9638967895507813,
-      "step": 4600
-    },
-    {
-      "epoch": 0.094,
-      "grad_norm": 0.36029258370399475,
-      "learning_rate": 0.001989442348614863,
-      "loss": 3.972870178222656,
-      "step": 4700
-    },
-    {
-      "epoch": 0.096,
-      "grad_norm": 0.26471713185310364,
-      "learning_rate": 0.0019884621851367075,
-      "loss": 3.959658508300781,
-      "step": 4800
-    },
-    {
-      "epoch": 0.098,
-      "grad_norm": 0.23001010715961456,
-      "learning_rate": 0.001987438783120401,
-      "loss": 3.9595660400390624,
-      "step": 4900
-    },
-    {
-      "epoch": 0.1,
-      "grad_norm": 0.23091939091682434,
-      "learning_rate": 0.001986372187332862,
-      "loss": 3.9660198974609373,
-      "step": 5000
-    },
-    {
-      "epoch": 0.1,
-      "eval_accuracy": 0.3069041095890411,
-      "eval_loss": 3.95263409614563,
-      "eval_runtime": 5.269,
-      "eval_samples_per_second": 189.788,
-      "eval_steps_per_second": 1.518,
-      "step": 5000
-    },
-    {
-      "epoch": 0.102,
-      "grad_norm": 0.28329843282699585,
-      "learning_rate": 0.0019852624444304467,
-      "loss": 3.9279766845703126,
-      "step": 5100
-    },
-    {
-      "epoch": 0.104,
-      "grad_norm": 0.4390493333339691,
-      "learning_rate": 0.0019841096029569044,
-      "loss": 3.940526123046875,
-      "step": 5200
-    },
-    {
-      "epoch": 0.106,
-      "grad_norm": 0.22625097632408142,
-      "learning_rate": 0.0019829137133412556,
-      "loss": 3.9342303466796875,
-      "step": 5300
-    },
-    {
-      "epoch": 0.108,
-      "grad_norm": 0.21681281924247742,
-      "learning_rate": 0.001981674827895587,
-      "loss": 3.9172735595703125,
-      "step": 5400
-    },
-    {
-      "epoch": 0.11,
-      "grad_norm": 0.3054690957069397,
-      "learning_rate": 0.00198039300081276,
-      "loss": 3.920041198730469,
-      "step": 5500
-    },
-    {
-      "epoch": 0.112,
-      "grad_norm": 0.24492567777633667,
-      "learning_rate": 0.0019790682881640448,
-      "loss": 3.918515319824219,
-      "step": 5600
-    },
-    {
-      "epoch": 0.114,
-      "grad_norm": 0.20615942776203156,
-      "learning_rate": 0.001977700747896664,
-      "loss": 3.902435302734375,
-      "step": 5700
-    },
-    {
-      "epoch": 0.116,
-      "grad_norm": 0.2998456358909607,
-      "learning_rate": 0.001976290439831259,
-      "loss": 3.9016647338867188,
-      "step": 5800
-    },
-    {
-      "epoch": 0.118,
-      "grad_norm": 0.22004957497119904,
-      "learning_rate": 0.0019748374256592736,
-      "loss": 3.897244567871094,
-      "step": 5900
-    },
-    {
-      "epoch": 0.12,
-      "grad_norm": 0.2709659934043884,
-      "learning_rate": 0.0019733417689402543,
-      "loss": 3.892081604003906,
-      "step": 6000
-    },
-    {
-      "epoch": 0.12,
-      "eval_accuracy": 0.31208414872798435,
-      "eval_loss": 3.8932552337646484,
-      "eval_runtime": 5.3089,
-      "eval_samples_per_second": 188.362,
-      "eval_steps_per_second": 1.507,
-      "step": 6000
     }
   ],
   "logging_steps": 100,
@@ -502,7 +344,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 128,
+  "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null
 }

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3ab34ea13b0e64af4c31a655731e444610b50d6a753b685ccf0d645cea5f76a3
-size 5265
+oid sha256:360066b4a5f1cc8c89941d389baaf9061071c4da2a624ff0fb9b3a9211c62ff7
+size 5905