Training in progress, step 16000, checkpoint

checkpoint-16000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:58e50bb8a5260aa6565a06fd28d95ed0aa393ab0e3403dd4b17ff6686e048eef
+oid sha256:00da54c3df53aa555e6eb10db490d10235f2ac5b3a2b3469edf0b514ff6dce3a
 size 1711848436

checkpoint-16000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7dc7629c5f432c5d1bf464c9caaf26d4d304135237f1d8a6049f85efee0ffa7b
+oid sha256:2704d43b2c9749de2bfae3d74d80f99fb7088122480292bc54a705719529b935
 size 3424043887

checkpoint-16000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:00e38b29b337c4a9e8f0159a9f973bcb2b561f1f32b5a9726c85faaac055b337
+oid sha256:3342d4b777ea71d9445bc1512c42ca6ec291e710630bf9d8d024a49c04b22fa0
 size 623

checkpoint-16000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.890542209148407,
+  "best_metric": 0.8807593584060669,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l58-l/checkpoint-16000",
-  "epoch": 2.947135752440597,
+  "epoch": 2.9476787030213707,
   "eval_steps": 500,
   "global_step": 16000,
   "is_hyper_param_search": false,
@@ -9,488 +9,488 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.09209799226376865,
-      "grad_norm": 0.0014697719598188996,
-      "learning_rate": 4.884877509670289e-08,
-      "loss": 0.4588,
+      "epoch": 0.09211495946941783,
+      "grad_norm": 20.28152847290039,
+      "learning_rate": 4.8848563006632274e-08,
+      "loss": 0.3758,
       "step": 500
     },
     {
-      "epoch": 0.09209799226376865,
-      "eval_loss": 1.4194883108139038,
-      "eval_runtime": 78.7703,
-      "eval_samples_per_second": 15.31,
-      "eval_steps_per_second": 1.917,
+      "epoch": 0.09211495946941783,
+      "eval_loss": 1.418367862701416,
+      "eval_runtime": 73.4439,
+      "eval_samples_per_second": 16.434,
+      "eval_steps_per_second": 2.056,
       "step": 500
     },
     {
-      "epoch": 0.1841959845275373,
-      "grad_norm": 301.76043701171875,
-      "learning_rate": 4.769755019340578e-08,
-      "loss": 0.4255,
+      "epoch": 0.18422991893883567,
+      "grad_norm": 47.29538345336914,
+      "learning_rate": 4.7697126013264556e-08,
+      "loss": 0.4104,
       "step": 1000
     },
     {
-      "epoch": 0.1841959845275373,
-      "eval_loss": 1.3416553735733032,
-      "eval_runtime": 77.0051,
-      "eval_samples_per_second": 15.661,
-      "eval_steps_per_second": 1.961,
+      "epoch": 0.18422991893883567,
+      "eval_loss": 1.3491536378860474,
+      "eval_runtime": 74.3404,
+      "eval_samples_per_second": 16.236,
+      "eval_steps_per_second": 2.031,
       "step": 1000
     },
     {
-      "epoch": 0.27629397679130596,
-      "grad_norm": 35.28853225708008,
-      "learning_rate": 4.654632529010867e-08,
-      "loss": 0.3724,
+      "epoch": 0.2763448784082535,
+      "grad_norm": 0.5450155138969421,
+      "learning_rate": 4.6545689019896826e-08,
+      "loss": 0.4325,
       "step": 1500
     },
     {
-      "epoch": 0.27629397679130596,
-      "eval_loss": 1.2872973680496216,
-      "eval_runtime": 77.5439,
-      "eval_samples_per_second": 15.552,
-      "eval_steps_per_second": 1.947,
+      "epoch": 0.2763448784082535,
+      "eval_loss": 1.2877168655395508,
+      "eval_runtime": 75.3303,
+      "eval_samples_per_second": 16.023,
+      "eval_steps_per_second": 2.005,
       "step": 1500
     },
     {
-      "epoch": 0.3683919690550746,
-      "grad_norm": 338.9349060058594,
-      "learning_rate": 4.539510038681156e-08,
-      "loss": 0.3251,
+      "epoch": 0.36845983787767134,
+      "grad_norm": 0.11449883133172989,
+      "learning_rate": 4.539425202652911e-08,
+      "loss": 0.3423,
       "step": 2000
     },
     {
-      "epoch": 0.3683919690550746,
-      "eval_loss": 1.234910249710083,
-      "eval_runtime": 78.7715,
-      "eval_samples_per_second": 15.31,
-      "eval_steps_per_second": 1.917,
+      "epoch": 0.36845983787767134,
+      "eval_loss": 1.2384389638900757,
+      "eval_runtime": 76.113,
+      "eval_samples_per_second": 15.858,
+      "eval_steps_per_second": 1.984,
       "step": 2000
     },
     {
-      "epoch": 0.46048996131884323,
-      "grad_norm": 325.85101318359375,
-      "learning_rate": 4.4243875483514457e-08,
-      "loss": 0.3308,
+      "epoch": 0.46057479734708917,
+      "grad_norm": 389.7380065917969,
+      "learning_rate": 4.4242815033161385e-08,
+      "loss": 0.3645,
       "step": 2500
     },
     {
-      "epoch": 0.46048996131884323,
-      "eval_loss": 1.1944907903671265,
-      "eval_runtime": 79.0397,
-      "eval_samples_per_second": 15.258,
-      "eval_steps_per_second": 1.91,
+      "epoch": 0.46057479734708917,
+      "eval_loss": 1.1896915435791016,
+      "eval_runtime": 76.23,
+      "eval_samples_per_second": 15.834,
+      "eval_steps_per_second": 1.981,
       "step": 2500
     },
     {
-      "epoch": 0.5525879535826119,
-      "grad_norm": 100.85855102539062,
-      "learning_rate": 4.309265058021735e-08,
-      "loss": 0.3017,
+      "epoch": 0.552689756816507,
+      "grad_norm": 340.1103820800781,
+      "learning_rate": 4.309137803979366e-08,
+      "loss": 0.3171,
       "step": 3000
     },
     {
-      "epoch": 0.5525879535826119,
-      "eval_loss": 1.1593009233474731,
-      "eval_runtime": 77.6354,
-      "eval_samples_per_second": 15.534,
-      "eval_steps_per_second": 1.945,
+      "epoch": 0.552689756816507,
+      "eval_loss": 1.149291753768921,
+      "eval_runtime": 76.1662,
+      "eval_samples_per_second": 15.847,
+      "eval_steps_per_second": 1.983,
       "step": 3000
     },
     {
-      "epoch": 0.6446859458463805,
-      "grad_norm": 8.865598678588867,
-      "learning_rate": 4.194142567692024e-08,
-      "loss": 0.2962,
+      "epoch": 0.6448047162859248,
+      "grad_norm": 196.7362518310547,
+      "learning_rate": 4.193994104642594e-08,
+      "loss": 0.275,
       "step": 3500
     },
     {
-      "epoch": 0.6446859458463805,
-      "eval_loss": 1.1259427070617676,
-      "eval_runtime": 78.2715,
-      "eval_samples_per_second": 15.408,
-      "eval_steps_per_second": 1.929,
+      "epoch": 0.6448047162859248,
+      "eval_loss": 1.1162768602371216,
+      "eval_runtime": 74.6569,
+      "eval_samples_per_second": 16.167,
+      "eval_steps_per_second": 2.023,
       "step": 3500
     },
     {
-      "epoch": 0.7367839381101492,
-      "grad_norm": 4.393447399139404,
-      "learning_rate": 4.079020077362314e-08,
-      "loss": 0.2919,
+      "epoch": 0.7369196757553427,
+      "grad_norm": 16.604211807250977,
+      "learning_rate": 4.078850405305821e-08,
+      "loss": 0.2907,
       "step": 4000
     },
     {
-      "epoch": 0.7367839381101492,
-      "eval_loss": 1.0954023599624634,
-      "eval_runtime": 78.5366,
-      "eval_samples_per_second": 15.356,
-      "eval_steps_per_second": 1.923,
+      "epoch": 0.7369196757553427,
+      "eval_loss": 1.091786503791809,
+      "eval_runtime": 76.3333,
+      "eval_samples_per_second": 15.812,
+      "eval_steps_per_second": 1.978,
       "step": 4000
     },
     {
-      "epoch": 0.8288819303739179,
-      "grad_norm": 1.6036031246185303,
-      "learning_rate": 3.9638975870326026e-08,
-      "loss": 0.307,
+      "epoch": 0.8290346352247605,
+      "grad_norm": 527.0987548828125,
+      "learning_rate": 3.9637067059690496e-08,
+      "loss": 0.2581,
       "step": 4500
     },
     {
-      "epoch": 0.8288819303739179,
-      "eval_loss": 1.0728861093521118,
-      "eval_runtime": 78.4032,
-      "eval_samples_per_second": 15.382,
-      "eval_steps_per_second": 1.926,
+      "epoch": 0.8290346352247605,
+      "eval_loss": 1.0729899406433105,
+      "eval_runtime": 74.4912,
+      "eval_samples_per_second": 16.203,
+      "eval_steps_per_second": 2.027,
       "step": 4500
     },
     {
-      "epoch": 0.9209799226376865,
-      "grad_norm": 11.990758895874023,
-      "learning_rate": 3.8487750967028915e-08,
-      "loss": 0.2764,
+      "epoch": 0.9211495946941783,
+      "grad_norm": 1.148181676864624,
+      "learning_rate": 3.848563006632277e-08,
+      "loss": 0.268,
       "step": 5000
     },
     {
-      "epoch": 0.9209799226376865,
-      "eval_loss": 1.0523593425750732,
-      "eval_runtime": 78.3726,
-      "eval_samples_per_second": 15.388,
-      "eval_steps_per_second": 1.927,
+      "epoch": 0.9211495946941783,
+      "eval_loss": 1.052911400794983,
+      "eval_runtime": 74.9586,
+      "eval_samples_per_second": 16.102,
+      "eval_steps_per_second": 2.014,
       "step": 5000
     },
     {
-      "epoch": 1.013077914901455,
-      "grad_norm": 43.865562438964844,
-      "learning_rate": 3.733652606373181e-08,
-      "loss": 0.2456,
+      "epoch": 1.013264554163596,
+      "grad_norm": 0.1680648922920227,
+      "learning_rate": 3.733419307295505e-08,
+      "loss": 0.2467,
       "step": 5500
     },
     {
-      "epoch": 1.013077914901455,
-      "eval_loss": 1.0375442504882812,
-      "eval_runtime": 78.2473,
-      "eval_samples_per_second": 15.413,
-      "eval_steps_per_second": 1.93,
+      "epoch": 1.013264554163596,
+      "eval_loss": 1.036412239074707,
+      "eval_runtime": 74.707,
+      "eval_samples_per_second": 16.156,
+      "eval_steps_per_second": 2.021,
       "step": 5500
     },
     {
-      "epoch": 1.1051759071652238,
-      "grad_norm": 105.91036987304688,
-      "learning_rate": 3.61853011604347e-08,
-      "loss": 0.2642,
+      "epoch": 1.105379513633014,
+      "grad_norm": 269.06585693359375,
+      "learning_rate": 3.6182756079587324e-08,
+      "loss": 0.2594,
       "step": 6000
     },
     {
-      "epoch": 1.1051759071652238,
-      "eval_loss": 1.0233356952667236,
-      "eval_runtime": 78.3135,
-      "eval_samples_per_second": 15.4,
-      "eval_steps_per_second": 1.928,
+      "epoch": 1.105379513633014,
+      "eval_loss": 1.0190415382385254,
+      "eval_runtime": 74.5657,
+      "eval_samples_per_second": 16.187,
+      "eval_steps_per_second": 2.025,
       "step": 6000
     },
     {
-      "epoch": 1.1972738994289924,
-      "grad_norm": 221.4896240234375,
-      "learning_rate": 3.503407625713759e-08,
-      "loss": 0.2066,
+      "epoch": 1.1974944731024317,
+      "grad_norm": 457.15850830078125,
+      "learning_rate": 3.50313190862196e-08,
+      "loss": 0.2153,
       "step": 6500
     },
     {
-      "epoch": 1.1972738994289924,
-      "eval_loss": 1.0104238986968994,
-      "eval_runtime": 78.6485,
-      "eval_samples_per_second": 15.334,
-      "eval_steps_per_second": 1.92,
+      "epoch": 1.1974944731024317,
+      "eval_loss": 1.0054280757904053,
+      "eval_runtime": 74.5269,
+      "eval_samples_per_second": 16.195,
+      "eval_steps_per_second": 2.026,
       "step": 6500
     },
     {
-      "epoch": 1.289371891692761,
-      "grad_norm": 270.53253173828125,
-      "learning_rate": 3.3882851353840485e-08,
-      "loss": 0.2376,
+      "epoch": 1.2896094325718497,
+      "grad_norm": 0.5831326842308044,
+      "learning_rate": 3.387988209285188e-08,
+      "loss": 0.1792,
       "step": 7000
     },
     {
-      "epoch": 1.289371891692761,
-      "eval_loss": 0.998353898525238,
-      "eval_runtime": 78.5502,
-      "eval_samples_per_second": 15.353,
-      "eval_steps_per_second": 1.922,
+      "epoch": 1.2896094325718497,
+      "eval_loss": 0.9938780665397644,
+      "eval_runtime": 74.5283,
+      "eval_samples_per_second": 16.195,
+      "eval_steps_per_second": 2.026,
       "step": 7000
     },
     {
-      "epoch": 1.3814698839565298,
-      "grad_norm": 402.9339904785156,
-      "learning_rate": 3.2731626450543374e-08,
-      "loss": 0.1931,
+      "epoch": 1.3817243920412676,
+      "grad_norm": 54.12948226928711,
+      "learning_rate": 3.272844509948415e-08,
+      "loss": 0.208,
       "step": 7500
     },
     {
-      "epoch": 1.3814698839565298,
-      "eval_loss": 0.9887453317642212,
-      "eval_runtime": 78.4833,
-      "eval_samples_per_second": 15.366,
-      "eval_steps_per_second": 1.924,
+      "epoch": 1.3817243920412676,
+      "eval_loss": 0.9816781282424927,
+      "eval_runtime": 76.244,
+      "eval_samples_per_second": 15.831,
+      "eval_steps_per_second": 1.98,
       "step": 7500
     },
     {
-      "epoch": 1.4735678762202984,
-      "grad_norm": 0.006355441175401211,
-      "learning_rate": 3.158040154724626e-08,
-      "loss": 0.2163,
+      "epoch": 1.4738393515106853,
+      "grad_norm": 0.09778845310211182,
+      "learning_rate": 3.1577008106116435e-08,
+      "loss": 0.1955,
       "step": 8000
     },
     {
-      "epoch": 1.4735678762202984,
-      "eval_loss": 0.9767189621925354,
-      "eval_runtime": 78.5499,
-      "eval_samples_per_second": 15.353,
-      "eval_steps_per_second": 1.922,
+      "epoch": 1.4738393515106853,
+      "eval_loss": 0.9722194671630859,
+      "eval_runtime": 74.4255,
+      "eval_samples_per_second": 16.218,
+      "eval_steps_per_second": 2.029,
       "step": 8000
     },
     {
-      "epoch": 1.565665868484067,
-      "grad_norm": 0.12982100248336792,
-      "learning_rate": 3.042917664394916e-08,
-      "loss": 0.1903,
+      "epoch": 1.565954310980103,
+      "grad_norm": 43.05193328857422,
+      "learning_rate": 3.042557111274871e-08,
+      "loss": 0.1943,
       "step": 8500
     },
     {
-      "epoch": 1.565665868484067,
-      "eval_loss": 0.9664921760559082,
-      "eval_runtime": 78.6832,
-      "eval_samples_per_second": 15.327,
-      "eval_steps_per_second": 1.919,
+      "epoch": 1.565954310980103,
+      "eval_loss": 0.962623119354248,
+      "eval_runtime": 74.5615,
+      "eval_samples_per_second": 16.188,
+      "eval_steps_per_second": 2.025,
       "step": 8500
     },
     {
-      "epoch": 1.6577638607478358,
-      "grad_norm": 173.31788635253906,
-      "learning_rate": 2.927795174065205e-08,
-      "loss": 0.2069,
+      "epoch": 1.658069270449521,
+      "grad_norm": 0.050531528890132904,
+      "learning_rate": 2.927413411938099e-08,
+      "loss": 0.2421,
       "step": 9000
     },
     {
-      "epoch": 1.6577638607478358,
-      "eval_loss": 0.9571623206138611,
-      "eval_runtime": 78.544,
-      "eval_samples_per_second": 15.354,
-      "eval_steps_per_second": 1.922,
+      "epoch": 1.658069270449521,
+      "eval_loss": 0.9543817043304443,
+      "eval_runtime": 76.2957,
+      "eval_samples_per_second": 15.82,
+      "eval_steps_per_second": 1.979,
       "step": 9000
     },
     {
-      "epoch": 1.7498618530116044,
-      "grad_norm": 358.7403259277344,
-      "learning_rate": 2.8126726837354947e-08,
-      "loss": 0.2093,
+      "epoch": 1.750184229918939,
+      "grad_norm": 0.07409189641475677,
+      "learning_rate": 2.8122697126013263e-08,
+      "loss": 0.2109,
       "step": 9500
     },
     {
-      "epoch": 1.7498618530116044,
-      "eval_loss": 0.9496769905090332,
-      "eval_runtime": 78.3124,
-      "eval_samples_per_second": 15.4,
-      "eval_steps_per_second": 1.928,
+      "epoch": 1.750184229918939,
+      "eval_loss": 0.9468172192573547,
+      "eval_runtime": 74.551,
+      "eval_samples_per_second": 16.19,
+      "eval_steps_per_second": 2.025,
       "step": 9500
     },
     {
-      "epoch": 1.841959845275373,
-      "grad_norm": 435.9449768066406,
-      "learning_rate": 2.6975501934057836e-08,
-      "loss": 0.2523,
+      "epoch": 1.8422991893883567,
+      "grad_norm": 1.4242560863494873,
+      "learning_rate": 2.697126013264554e-08,
+      "loss": 0.1799,
       "step": 10000
     },
     {
-      "epoch": 1.841959845275373,
-      "eval_loss": 0.9420022368431091,
-      "eval_runtime": 78.3702,
-      "eval_samples_per_second": 15.388,
-      "eval_steps_per_second": 1.927,
+      "epoch": 1.8422991893883567,
+      "eval_loss": 0.9402961730957031,
+      "eval_runtime": 74.4464,
+      "eval_samples_per_second": 16.213,
+      "eval_steps_per_second": 2.028,
       "step": 10000
     },
     {
-      "epoch": 1.9340578375391417,
-      "grad_norm": 23.49118423461914,
-      "learning_rate": 2.582427703076073e-08,
-      "loss": 0.2127,
+      "epoch": 1.9344141488577744,
+      "grad_norm": 0.06974973529577255,
+      "learning_rate": 2.5819823139277818e-08,
+      "loss": 0.2114,
       "step": 10500
     },
     {
-      "epoch": 1.9340578375391417,
-      "eval_loss": 0.9328628182411194,
-      "eval_runtime": 76.6904,
-      "eval_samples_per_second": 15.726,
-      "eval_steps_per_second": 1.969,
+      "epoch": 1.9344141488577744,
+      "eval_loss": 0.9320051074028015,
+      "eval_runtime": 76.0983,
+      "eval_samples_per_second": 15.861,
+      "eval_steps_per_second": 1.984,
       "step": 10500
     },
     {
-      "epoch": 2.02615582980291,
-      "grad_norm": 0.4029064476490021,
-      "learning_rate": 2.467305212746362e-08,
-      "loss": 0.1968,
+      "epoch": 2.026529108327192,
+      "grad_norm": 0.17525330185890198,
+      "learning_rate": 2.4668386145910094e-08,
+      "loss": 0.2058,
       "step": 11000
     },
     {
-      "epoch": 2.02615582980291,
-      "eval_loss": 0.926975429058075,
-      "eval_runtime": 76.7161,
-      "eval_samples_per_second": 15.72,
-      "eval_steps_per_second": 1.968,
+      "epoch": 2.026529108327192,
+      "eval_loss": 0.9245836138725281,
+      "eval_runtime": 74.5035,
+      "eval_samples_per_second": 16.201,
+      "eval_steps_per_second": 2.027,
       "step": 11000
     },
     {
-      "epoch": 2.118253822066679,
-      "grad_norm": 274.0943603515625,
-      "learning_rate": 2.3521827224166513e-08,
-      "loss": 0.1879,
+      "epoch": 2.1186440677966103,
+      "grad_norm": 9.384895324707031,
+      "learning_rate": 2.3516949152542374e-08,
+      "loss": 0.188,
       "step": 11500
     },
     {
-      "epoch": 2.118253822066679,
-      "eval_loss": 0.9231382012367249,
-      "eval_runtime": 78.6224,
-      "eval_samples_per_second": 15.339,
-      "eval_steps_per_second": 1.921,
+      "epoch": 2.1186440677966103,
+      "eval_loss": 0.9178166389465332,
+      "eval_runtime": 74.4204,
+      "eval_samples_per_second": 16.219,
+      "eval_steps_per_second": 2.029,
       "step": 11500
     },
     {
-      "epoch": 2.2103518143304477,
-      "grad_norm": 0.000765918695833534,
-      "learning_rate": 2.2370602320869402e-08,
-      "loss": 0.1981,
+      "epoch": 2.210759027266028,
+      "grad_norm": 0.770260214805603,
+      "learning_rate": 2.236551215917465e-08,
+      "loss": 0.1657,
       "step": 12000
     },
     {
-      "epoch": 2.2103518143304477,
-      "eval_loss": 0.9183884263038635,
-      "eval_runtime": 76.7707,
-      "eval_samples_per_second": 15.709,
-      "eval_steps_per_second": 1.967,
+      "epoch": 2.210759027266028,
+      "eval_loss": 0.9122205376625061,
+      "eval_runtime": 76.1185,
+      "eval_samples_per_second": 15.857,
+      "eval_steps_per_second": 1.984,
       "step": 12000
     },
     {
-      "epoch": 2.302449806594216,
-      "grad_norm": 0.6250036358833313,
-      "learning_rate": 2.1219377417572295e-08,
-      "loss": 0.1964,
+      "epoch": 2.3028739867354457,
+      "grad_norm": 614.7385864257812,
+      "learning_rate": 2.1214075165806926e-08,
+      "loss": 0.1948,
       "step": 12500
     },
     {
-      "epoch": 2.302449806594216,
-      "eval_loss": 0.9134928584098816,
-      "eval_runtime": 78.5128,
-      "eval_samples_per_second": 15.361,
-      "eval_steps_per_second": 1.923,
+      "epoch": 2.3028739867354457,
+      "eval_loss": 0.90776127576828,
+      "eval_runtime": 74.5589,
+      "eval_samples_per_second": 16.189,
+      "eval_steps_per_second": 2.025,
       "step": 12500
     },
     {
-      "epoch": 2.394547798857985,
-      "grad_norm": 158.6990203857422,
-      "learning_rate": 2.0068152514275187e-08,
-      "loss": 0.1697,
+      "epoch": 2.3949889462048635,
+      "grad_norm": 0.7871919274330139,
+      "learning_rate": 2.0062638172439202e-08,
+      "loss": 0.1882,
       "step": 13000
     },
     {
-      "epoch": 2.394547798857985,
-      "eval_loss": 0.910048246383667,
-      "eval_runtime": 76.7476,
-      "eval_samples_per_second": 15.714,
-      "eval_steps_per_second": 1.967,
+      "epoch": 2.3949889462048635,
+      "eval_loss": 0.9029939770698547,
+      "eval_runtime": 74.3684,
+      "eval_samples_per_second": 16.23,
+      "eval_steps_per_second": 2.03,
       "step": 13000
     },
     {
-      "epoch": 2.4866457911217537,
-      "grad_norm": 1.288072943687439,
-      "learning_rate": 1.891692761097808e-08,
-      "loss": 0.2015,
+      "epoch": 2.4871039056742816,
+      "grad_norm": 1.3432903289794922,
+      "learning_rate": 1.891120117907148e-08,
+      "loss": 0.1874,
       "step": 13500
     },
     {
-      "epoch": 2.4866457911217537,
-      "eval_loss": 0.9051916599273682,
-      "eval_runtime": 78.3909,
-      "eval_samples_per_second": 15.384,
-      "eval_steps_per_second": 1.926,
+      "epoch": 2.4871039056742816,
+      "eval_loss": 0.8971672654151917,
+      "eval_runtime": 76.1444,
+      "eval_samples_per_second": 15.851,
+      "eval_steps_per_second": 1.983,
       "step": 13500
     },
     {
-      "epoch": 2.578743783385522,
-      "grad_norm": 5.835799311171286e-05,
-      "learning_rate": 1.7765702707680972e-08,
-      "loss": 0.1827,
+      "epoch": 2.5792188651436994,
+      "grad_norm": 343.2808532714844,
+      "learning_rate": 1.7759764185703757e-08,
+      "loss": 0.1423,
       "step": 14000
     },
     {
-      "epoch": 2.578743783385522,
-      "eval_loss": 0.9026296734809875,
-      "eval_runtime": 76.6591,
-      "eval_samples_per_second": 15.732,
-      "eval_steps_per_second": 1.97,
+      "epoch": 2.5792188651436994,
+      "eval_loss": 0.8923307657241821,
+      "eval_runtime": 74.6823,
+      "eval_samples_per_second": 16.162,
+      "eval_steps_per_second": 2.022,
       "step": 14000
     },
     {
-      "epoch": 2.670841775649291,
-      "grad_norm": 30.217784881591797,
-      "learning_rate": 1.6614477804383865e-08,
-      "loss": 0.1435,
+      "epoch": 2.671333824613117,
+      "grad_norm": 0.01757567748427391,
+      "learning_rate": 1.6608327192336037e-08,
+      "loss": 0.1424,
       "step": 14500
     },
     {
-      "epoch": 2.670841775649291,
-      "eval_loss": 0.899847686290741,
-      "eval_runtime": 76.742,
-      "eval_samples_per_second": 15.715,
-      "eval_steps_per_second": 1.968,
+      "epoch": 2.671333824613117,
+      "eval_loss": 0.888923704624176,
+      "eval_runtime": 74.5202,
+      "eval_samples_per_second": 16.197,
+      "eval_steps_per_second": 2.026,
       "step": 14500
     },
     {
-      "epoch": 2.7629397679130596,
-      "grad_norm": 202.30496215820312,
-      "learning_rate": 1.5463252901086757e-08,
-      "loss": 0.1541,
+      "epoch": 2.7634487840825352,
+      "grad_norm": 1.6630617380142212,
+      "learning_rate": 1.5456890198968313e-08,
+      "loss": 0.18,
       "step": 15000
     },
     {
-      "epoch": 2.7629397679130596,
-      "eval_loss": 0.896297812461853,
-      "eval_runtime": 78.5539,
-      "eval_samples_per_second": 15.353,
-      "eval_steps_per_second": 1.922,
+      "epoch": 2.7634487840825352,
+      "eval_loss": 0.885770320892334,
+      "eval_runtime": 76.123,
+      "eval_samples_per_second": 15.856,
+      "eval_steps_per_second": 1.984,
       "step": 15000
     },
     {
-      "epoch": 2.8550377601768284,
-      "grad_norm": 241.57069396972656,
-      "learning_rate": 1.4312027997789646e-08,
-      "loss": 0.1716,
+      "epoch": 2.855563743551953,
+      "grad_norm": 0.0005675083957612514,
+      "learning_rate": 1.430545320560059e-08,
+      "loss": 0.1699,
       "step": 15500
     },
     {
-      "epoch": 2.8550377601768284,
-      "eval_loss": 0.8935310244560242,
-      "eval_runtime": 76.9195,
-      "eval_samples_per_second": 15.679,
-      "eval_steps_per_second": 1.963,
+      "epoch": 2.855563743551953,
+      "eval_loss": 0.8834386467933655,
+      "eval_runtime": 74.4405,
+      "eval_samples_per_second": 16.214,
+      "eval_steps_per_second": 2.028,
       "step": 15500
     },
     {
-      "epoch": 2.947135752440597,
-      "grad_norm": 0.25739413499832153,
-      "learning_rate": 1.316080309449254e-08,
-      "loss": 0.2056,
+      "epoch": 2.9476787030213707,
+      "grad_norm": 232.96043395996094,
+      "learning_rate": 1.3154016212232866e-08,
+      "loss": 0.1711,
       "step": 16000
     },
     {
-      "epoch": 2.947135752440597,
-      "eval_loss": 0.890542209148407,
-      "eval_runtime": 76.7268,
-      "eval_samples_per_second": 15.718,
-      "eval_steps_per_second": 1.968,
+      "epoch": 2.9476787030213707,
+      "eval_loss": 0.8807593584060669,
+      "eval_runtime": 74.4453,
+      "eval_samples_per_second": 16.213,
+      "eval_steps_per_second": 2.028,
       "step": 16000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 21716,
+  "max_steps": 21712,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 4,
   "save_steps": 500,
@@ -506,7 +506,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5756493378511260.0,
+  "total_flos": 5756853136862520.0,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null

checkpoint-16000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:417b0bacc3a8958dfa75621764230fb890db48c09ca241dff24e5264d406285b
+oid sha256:595bb6f39a89ec7030e395471c015665757cb5cacaafa6bba1613fd32b629686
 size 4847