🍻 cheers

all_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 0.710220593319964,
-    "eval_loss": 0.9962999820709229,
-    "eval_runtime": 152.3809,
-    "eval_samples_per_second": 189.801,
-    "eval_steps_per_second": 23.73,
-    "train_loss": 0.8206716175314105,
-    "train_runtime": 13827.0136,
-    "train_samples_per_second": 66.935,
-    "train_steps_per_second": 2.092
+    "eval_accuracy": 0.7164788050618907,
+    "eval_loss": 0.9746271967887878,
+    "eval_runtime": 155.7191,
+    "eval_samples_per_second": 185.732,
+    "eval_steps_per_second": 23.221,
+    "train_loss": 0.9355616505553886,
+    "train_runtime": 13547.1303,
+    "train_samples_per_second": 68.318,
+    "train_steps_per_second": 2.135
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 0.710220593319964,
-    "eval_loss": 0.9962999820709229,
-    "eval_runtime": 152.3809,
-    "eval_samples_per_second": 189.801,
-    "eval_steps_per_second": 23.73
+    "eval_accuracy": 0.7164788050618907,
+    "eval_loss": 0.9746271967887878,
+    "eval_runtime": 155.7191,
+    "eval_samples_per_second": 185.732,
+    "eval_steps_per_second": 23.221
 }

runs/Jun06_14-38-53_adrian-development/events.out.tfevents.1686068860.adrian-development.77015.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ece15a64ee14ac17d8203a4835811612d3769d140597e75e275c59e1b1757c9d
+size 369

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 4.0,
-    "train_loss": 0.8206716175314105,
-    "train_runtime": 13827.0136,
-    "train_samples_per_second": 66.935,
-    "train_steps_per_second": 2.092
+    "train_loss": 0.9355616505553886,
+    "train_runtime": 13547.1303,
+    "train_samples_per_second": 68.318,
+    "train_steps_per_second": 2.135
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.9962999820709229,
-  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-14000",
+  "best_metric": 0.9746271967887878,
+  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-21000",
   "epoch": 4.0,
   "global_step": 28924,
   "is_hyper_param_search": false,
@@ -9,432 +9,432 @@
   "log_history": [
     {
       "epoch": 0.14,
-      "learning_rate": 7.240959065136219e-05,
-      "loss": 1.6085,
+      "learning_rate": 2.4135665883003737e-05,
+      "loss": 2.0439,
       "step": 1000
     },
     {
       "epoch": 0.14,
-      "eval_accuracy": 0.6414494156697323,
-      "eval_loss": 1.2965176105499268,
-      "eval_runtime": 155.6021,
-      "eval_samples_per_second": 185.872,
-      "eval_steps_per_second": 23.239,
+      "eval_accuracy": 0.6367125371689372,
+      "eval_loss": 1.4945471286773682,
+      "eval_runtime": 155.4022,
+      "eval_samples_per_second": 186.111,
+      "eval_steps_per_second": 23.269,
       "step": 1000
     },
     {
       "epoch": 0.28,
-      "learning_rate": 6.981918130272437e-05,
-      "loss": 1.2071,
+      "learning_rate": 2.327133176600747e-05,
+      "loss": 1.3784,
       "step": 2000
     },
     {
       "epoch": 0.28,
-      "eval_accuracy": 0.6689717170320172,
-      "eval_loss": 1.163772702217102,
-      "eval_runtime": 147.9377,
-      "eval_samples_per_second": 195.501,
-      "eval_steps_per_second": 24.443,
+      "eval_accuracy": 0.6662748081045571,
+      "eval_loss": 1.2658541202545166,
+      "eval_runtime": 152.9449,
+      "eval_samples_per_second": 189.101,
+      "eval_steps_per_second": 23.642,
       "step": 2000
     },
     {
       "epoch": 0.41,
-      "learning_rate": 6.722617895173557e-05,
-      "loss": 1.1467,
+      "learning_rate": 2.2408726317245195e-05,
+      "loss": 1.2223,
       "step": 3000
     },
     {
       "epoch": 0.41,
-      "eval_accuracy": 0.672187262291681,
-      "eval_loss": 1.1356933116912842,
-      "eval_runtime": 148.7126,
-      "eval_samples_per_second": 194.482,
-      "eval_steps_per_second": 24.315,
+      "eval_accuracy": 0.6780305649678445,
+      "eval_loss": 1.176424503326416,
+      "eval_runtime": 155.5666,
+      "eval_samples_per_second": 185.914,
+      "eval_steps_per_second": 23.244,
       "step": 3000
     },
     {
       "epoch": 0.55,
-      "learning_rate": 6.463836260544876e-05,
-      "loss": 1.1073,
+      "learning_rate": 2.1544392200248927e-05,
+      "loss": 1.153,
       "step": 4000
     },
     {
       "epoch": 0.55,
-      "eval_accuracy": 0.6832515040453634,
-      "eval_loss": 1.094034194946289,
-      "eval_runtime": 147.3532,
-      "eval_samples_per_second": 196.277,
-      "eval_steps_per_second": 24.54,
+      "eval_accuracy": 0.687227715925593,
+      "eval_loss": 1.115257740020752,
+      "eval_runtime": 154.7152,
+      "eval_samples_per_second": 186.937,
+      "eval_steps_per_second": 23.372,
       "step": 4000
     },
     {
       "epoch": 0.69,
-      "learning_rate": 6.204536025445996e-05,
-      "loss": 1.0721,
+      "learning_rate": 2.0680058083252663e-05,
+      "loss": 1.1037,
       "step": 5000
     },
     {
       "epoch": 0.69,
-      "eval_accuracy": 0.685879261461863,
-      "eval_loss": 1.0801600217819214,
-      "eval_runtime": 148.8743,
-      "eval_samples_per_second": 194.271,
-      "eval_steps_per_second": 24.289,
+      "eval_accuracy": 0.6912385035613028,
+      "eval_loss": 1.0900990962982178,
+      "eval_runtime": 154.22,
+      "eval_samples_per_second": 187.537,
+      "eval_steps_per_second": 23.447,
       "step": 5000
     },
     {
       "epoch": 0.83,
-      "learning_rate": 5.9452357903471155e-05,
-      "loss": 1.0607,
+      "learning_rate": 1.9816588300373394e-05,
+      "loss": 1.0827,
       "step": 6000
     },
     {
       "epoch": 0.83,
-      "eval_accuracy": 0.6946615033538482,
-      "eval_loss": 1.0509121417999268,
-      "eval_runtime": 153.6507,
-      "eval_samples_per_second": 188.232,
-      "eval_steps_per_second": 23.534,
+      "eval_accuracy": 0.6933822004010788,
+      "eval_loss": 1.068312406539917,
+      "eval_runtime": 150.0826,
+      "eval_samples_per_second": 192.707,
+      "eval_steps_per_second": 24.093,
       "step": 6000
     },
     {
       "epoch": 0.97,
-      "learning_rate": 5.685935555248236e-05,
-      "loss": 1.032,
+      "learning_rate": 1.8953118517494124e-05,
+      "loss": 1.0646,
       "step": 7000
     },
     {
       "epoch": 0.97,
-      "eval_accuracy": 0.6914805338496646,
-      "eval_loss": 1.055686593055725,
-      "eval_runtime": 151.6461,
-      "eval_samples_per_second": 190.72,
-      "eval_steps_per_second": 23.845,
+      "eval_accuracy": 0.7026485028697877,
+      "eval_loss": 1.037534475326538,
+      "eval_runtime": 148.9104,
+      "eval_samples_per_second": 194.224,
+      "eval_steps_per_second": 24.283,
       "step": 7000
     },
     {
       "epoch": 1.11,
-      "learning_rate": 5.4266353201493564e-05,
-      "loss": 0.9224,
+      "learning_rate": 1.8088784400497857e-05,
+      "loss": 0.9782,
       "step": 8000
     },
     {
       "epoch": 1.11,
-      "eval_accuracy": 0.696632321416223,
-      "eval_loss": 1.0505746603012085,
-      "eval_runtime": 152.4908,
-      "eval_samples_per_second": 189.664,
-      "eval_steps_per_second": 23.713,
+      "eval_accuracy": 0.6986722909895581,
+      "eval_loss": 1.0466938018798828,
+      "eval_runtime": 149.1228,
+      "eval_samples_per_second": 193.948,
+      "eval_steps_per_second": 24.248,
       "step": 8000
     },
     {
       "epoch": 1.24,
-      "learning_rate": 5.167335085050477e-05,
-      "loss": 0.9029,
+      "learning_rate": 1.7225314617618587e-05,
+      "loss": 0.9627,
       "step": 9000
     },
     {
       "epoch": 1.24,
-      "eval_accuracy": 0.6952492911970126,
-      "eval_loss": 1.0421319007873535,
-      "eval_runtime": 153.6492,
-      "eval_samples_per_second": 188.234,
-      "eval_steps_per_second": 23.534,
+      "eval_accuracy": 0.7060023511513727,
+      "eval_loss": 1.0178191661834717,
+      "eval_runtime": 149.6361,
+      "eval_samples_per_second": 193.282,
+      "eval_steps_per_second": 24.165,
       "step": 9000
     },
     {
       "epoch": 1.38,
-      "learning_rate": 4.9080348499515966e-05,
-      "loss": 0.8858,
+      "learning_rate": 1.6360980500622323e-05,
+      "loss": 0.9411,
       "step": 10000
     },
     {
       "epoch": 1.38,
-      "eval_accuracy": 0.701887836249222,
-      "eval_loss": 1.0203652381896973,
-      "eval_runtime": 153.0256,
-      "eval_samples_per_second": 189.001,
-      "eval_steps_per_second": 23.63,
+      "eval_accuracy": 0.7056911693520503,
+      "eval_loss": 1.0125459432601929,
+      "eval_runtime": 148.9456,
+      "eval_samples_per_second": 194.178,
+      "eval_steps_per_second": 24.277,
       "step": 10000
     },
     {
       "epoch": 1.52,
-      "learning_rate": 4.6489939150878154e-05,
-      "loss": 0.8943,
+      "learning_rate": 1.5496646383626052e-05,
+      "loss": 0.9333,
       "step": 11000
     },
     {
       "epoch": 1.52,
-      "eval_accuracy": 0.7037895028006362,
-      "eval_loss": 1.0182340145111084,
-      "eval_runtime": 153.2325,
-      "eval_samples_per_second": 188.746,
-      "eval_steps_per_second": 23.598,
+      "eval_accuracy": 0.7067630177719383,
+      "eval_loss": 1.0095747709274292,
+      "eval_runtime": 149.6511,
+      "eval_samples_per_second": 193.263,
+      "eval_steps_per_second": 24.163,
       "step": 11000
     },
     {
       "epoch": 1.66,
-      "learning_rate": 4.389952980224035e-05,
-      "loss": 0.8756,
+      "learning_rate": 1.4633176600746785e-05,
+      "loss": 0.9333,
       "step": 12000
     },
     {
       "epoch": 1.66,
-      "eval_accuracy": 0.7108083811631284,
-      "eval_loss": 1.0010631084442139,
-      "eval_runtime": 153.3911,
-      "eval_samples_per_second": 188.551,
-      "eval_steps_per_second": 23.574,
+      "eval_accuracy": 0.7047576239540834,
+      "eval_loss": 1.0087416172027588,
+      "eval_runtime": 149.4203,
+      "eval_samples_per_second": 193.561,
+      "eval_steps_per_second": 24.2,
       "step": 12000
     },
     {
       "epoch": 1.8,
-      "learning_rate": 4.130652745125155e-05,
-      "loss": 0.8657,
+      "learning_rate": 1.3768842483750519e-05,
+      "loss": 0.9227,
       "step": 13000
     },
     {
       "epoch": 1.8,
-      "eval_accuracy": 0.7074199571260632,
-      "eval_loss": 1.0035113096237183,
-      "eval_runtime": 150.8806,
-      "eval_samples_per_second": 191.688,
-      "eval_steps_per_second": 23.966,
+      "eval_accuracy": 0.7118456538275362,
+      "eval_loss": 0.9883840084075928,
+      "eval_runtime": 149.2315,
+      "eval_samples_per_second": 193.806,
+      "eval_steps_per_second": 24.231,
       "step": 13000
     },
     {
       "epoch": 1.94,
-      "learning_rate": 3.871352510026276e-05,
-      "loss": 0.8737,
+      "learning_rate": 1.290537270087125e-05,
+      "loss": 0.9131,
       "step": 14000
     },
     {
       "epoch": 1.94,
-      "eval_accuracy": 0.710220593319964,
-      "eval_loss": 0.9962999820709229,
-      "eval_runtime": 153.1018,
-      "eval_samples_per_second": 188.907,
-      "eval_steps_per_second": 23.618,
+      "eval_accuracy": 0.7127100477145426,
+      "eval_loss": 0.986507773399353,
+      "eval_runtime": 149.078,
+      "eval_samples_per_second": 194.006,
+      "eval_steps_per_second": 24.256,
       "step": 14000
     },
     {
       "epoch": 2.07,
-      "learning_rate": 3.6120522749273955e-05,
-      "loss": 0.7893,
+      "learning_rate": 1.2041038583874984e-05,
+      "loss": 0.8703,
       "step": 15000
     },
     {
       "epoch": 2.07,
-      "eval_accuracy": 0.708872138856234,
-      "eval_loss": 1.020798921585083,
-      "eval_runtime": 154.3576,
-      "eval_samples_per_second": 187.37,
-      "eval_steps_per_second": 23.426,
+      "eval_accuracy": 0.7143696839775949,
+      "eval_loss": 0.9831659197807312,
+      "eval_runtime": 150.612,
+      "eval_samples_per_second": 192.03,
+      "eval_steps_per_second": 24.009,
       "step": 15000
     },
     {
       "epoch": 2.21,
-      "learning_rate": 3.352752039828516e-05,
-      "loss": 0.7067,
+      "learning_rate": 1.1177568800995714e-05,
+      "loss": 0.8064,
       "step": 16000
     },
     {
       "epoch": 2.21,
-      "eval_accuracy": 0.7075582601479843,
-      "eval_loss": 1.0219132900238037,
-      "eval_runtime": 152.5343,
-      "eval_samples_per_second": 189.61,
-      "eval_steps_per_second": 23.706,
+      "eval_accuracy": 0.7136435931125095,
+      "eval_loss": 0.9858880043029785,
+      "eval_runtime": 149.8659,
+      "eval_samples_per_second": 192.986,
+      "eval_steps_per_second": 24.128,
       "step": 16000
     },
     {
       "epoch": 2.35,
-      "learning_rate": 3.093451804729636e-05,
-      "loss": 0.7072,
+      "learning_rate": 1.0313234683999447e-05,
+      "loss": 0.8131,
       "step": 17000
     },
     {
       "epoch": 2.35,
-      "eval_accuracy": 0.7095982297213194,
-      "eval_loss": 1.0180853605270386,
-      "eval_runtime": 154.4198,
-      "eval_samples_per_second": 187.295,
-      "eval_steps_per_second": 23.417,
+      "eval_accuracy": 0.7127100477145426,
+      "eval_loss": 0.9879944920539856,
+      "eval_runtime": 150.5293,
+      "eval_samples_per_second": 192.135,
+      "eval_steps_per_second": 24.022,
       "step": 17000
     },
     {
       "epoch": 2.49,
-      "learning_rate": 2.834410869865855e-05,
-      "loss": 0.6914,
+      "learning_rate": 9.448900567003181e-06,
+      "loss": 0.8064,
       "step": 18000
     },
     {
       "epoch": 2.49,
-      "eval_accuracy": 0.7122605628932992,
-      "eval_loss": 1.0164895057678223,
-      "eval_runtime": 151.314,
-      "eval_samples_per_second": 191.139,
-      "eval_steps_per_second": 23.897,
+      "eval_accuracy": 0.7113270174953322,
+      "eval_loss": 0.9872538447380066,
+      "eval_runtime": 149.9196,
+      "eval_samples_per_second": 192.917,
+      "eval_steps_per_second": 24.12,
       "step": 18000
     },
     {
       "epoch": 2.63,
-      "learning_rate": 2.575369935002074e-05,
-      "loss": 0.7044,
+      "learning_rate": 8.584566450006915e-06,
+      "loss": 0.812,
       "step": 19000
     },
     {
       "epoch": 2.63,
-      "eval_accuracy": 0.71236429015974,
-      "eval_loss": 1.01729416847229,
-      "eval_runtime": 151.1469,
-      "eval_samples_per_second": 191.35,
-      "eval_steps_per_second": 23.924,
+      "eval_accuracy": 0.7153723808865223,
+      "eval_loss": 0.9793804883956909,
+      "eval_runtime": 151.3957,
+      "eval_samples_per_second": 191.036,
+      "eval_steps_per_second": 23.884,
       "step": 19000
     },
     {
       "epoch": 2.77,
-      "learning_rate": 2.3160696999031945e-05,
-      "loss": 0.7014,
+      "learning_rate": 7.721096667127644e-06,
+      "loss": 0.7977,
       "step": 20000
     },
     {
       "epoch": 2.77,
-      "eval_accuracy": 0.7145425627549962,
-      "eval_loss": 1.0055793523788452,
-      "eval_runtime": 153.0154,
-      "eval_samples_per_second": 189.014,
-      "eval_steps_per_second": 23.632,
+      "eval_accuracy": 0.7151649263536408,
+      "eval_loss": 0.9755117297172546,
+      "eval_runtime": 144.4577,
+      "eval_samples_per_second": 200.211,
+      "eval_steps_per_second": 25.032,
       "step": 20000
     },
     {
       "epoch": 2.9,
-      "learning_rate": 2.0567694648043142e-05,
-      "loss": 0.6997,
+      "learning_rate": 6.856762550131379e-06,
+      "loss": 0.8118,
       "step": 21000
     },
     {
       "epoch": 2.9,
-      "eval_accuracy": 0.7116381992946545,
-      "eval_loss": 1.0049206018447876,
-      "eval_runtime": 154.2346,
-      "eval_samples_per_second": 187.52,
-      "eval_steps_per_second": 23.445,
+      "eval_accuracy": 0.7164788050618907,
+      "eval_loss": 0.9746271967887878,
+      "eval_runtime": 155.3329,
+      "eval_samples_per_second": 186.194,
+      "eval_steps_per_second": 23.279,
       "step": 21000
     },
     {
       "epoch": 3.04,
-      "learning_rate": 1.7974692297054347e-05,
-      "loss": 0.6378,
+      "learning_rate": 5.992428433135113e-06,
+      "loss": 0.7741,
       "step": 22000
     },
     {
       "epoch": 3.04,
-      "eval_accuracy": 0.7104971993638061,
-      "eval_loss": 1.035260558128357,
-      "eval_runtime": 154.7462,
-      "eval_samples_per_second": 186.9,
-      "eval_steps_per_second": 23.367,
+      "eval_accuracy": 0.7189336837009889,
+      "eval_loss": 0.9803725481033325,
+      "eval_runtime": 153.5991,
+      "eval_samples_per_second": 188.295,
+      "eval_steps_per_second": 23.542,
       "step": 22000
     },
     {
       "epoch": 3.18,
-      "learning_rate": 1.538168994606555e-05,
-      "loss": 0.5446,
+      "learning_rate": 5.128958650255843e-06,
+      "loss": 0.7202,
       "step": 23000
     },
     {
       "epoch": 3.18,
-      "eval_accuracy": 0.7085609570569117,
-      "eval_loss": 1.0574305057525635,
-      "eval_runtime": 153.225,
-      "eval_samples_per_second": 188.755,
-      "eval_steps_per_second": 23.599,
+      "eval_accuracy": 0.71371274462347,
+      "eval_loss": 0.9932627081871033,
+      "eval_runtime": 153.8948,
+      "eval_samples_per_second": 187.934,
+      "eval_steps_per_second": 23.497,
       "step": 23000
     },
     {
       "epoch": 3.32,
-      "learning_rate": 1.2788687595076752e-05,
-      "loss": 0.5307,
+      "learning_rate": 4.264624533259577e-06,
+      "loss": 0.7186,
       "step": 24000
     },
     {
       "epoch": 3.32,
-      "eval_accuracy": 0.7079040177027868,
-      "eval_loss": 1.0585498809814453,
-      "eval_runtime": 150.418,
-      "eval_samples_per_second": 192.277,
-      "eval_steps_per_second": 24.04,
+      "eval_accuracy": 0.7173086231934168,
+      "eval_loss": 0.9890537261962891,
+      "eval_runtime": 154.048,
+      "eval_samples_per_second": 187.747,
+      "eval_steps_per_second": 23.473,
       "step": 24000
     },
     {
       "epoch": 3.46,
-      "learning_rate": 1.0198278246438942e-05,
-      "loss": 0.5269,
+      "learning_rate": 3.4011547503803073e-06,
+      "loss": 0.7164,
       "step": 25000
     },
     {
       "epoch": 3.46,
-      "eval_accuracy": 0.7093907751884378,
-      "eval_loss": 1.0660898685455322,
-      "eval_runtime": 153.521,
-      "eval_samples_per_second": 188.391,
-      "eval_steps_per_second": 23.554,
+      "eval_accuracy": 0.7176198049927391,
+      "eval_loss": 0.9812184572219849,
+      "eval_runtime": 153.9568,
+      "eval_samples_per_second": 187.858,
+      "eval_steps_per_second": 23.487,
       "step": 25000
     },
     {
       "epoch": 3.6,
-      "learning_rate": 7.607868897801134e-06,
-      "loss": 0.525,
+      "learning_rate": 2.536820633384041e-06,
+      "loss": 0.7047,
       "step": 26000
     },
     {
       "epoch": 3.6,
-      "eval_accuracy": 0.7103934720973654,
-      "eval_loss": 1.059892177581787,
-      "eval_runtime": 153.2046,
-      "eval_samples_per_second": 188.78,
-      "eval_steps_per_second": 23.602,
+      "eval_accuracy": 0.7157527141968052,
+      "eval_loss": 0.9885823130607605,
+      "eval_runtime": 155.4293,
+      "eval_samples_per_second": 186.078,
+      "eval_steps_per_second": 23.265,
       "step": 26000
     },
     {
       "epoch": 3.73,
-      "learning_rate": 5.0148665468123355e-06,
-      "loss": 0.516,
+      "learning_rate": 1.6724865163877752e-06,
+      "loss": 0.7241,
       "step": 27000
     },
     {
       "epoch": 3.73,
-      "eval_accuracy": 0.7111195629624507,
-      "eval_loss": 1.0657610893249512,
-      "eval_runtime": 152.7663,
-      "eval_samples_per_second": 189.322,
-      "eval_steps_per_second": 23.67,
+      "eval_accuracy": 0.7153032293755619,
+      "eval_loss": 0.9851787686347961,
+      "eval_runtime": 155.7881,
+      "eval_samples_per_second": 185.65,
+      "eval_steps_per_second": 23.211,
       "step": 27000
     },
     {
       "epoch": 3.87,
-      "learning_rate": 2.4244571981745263e-06,
-      "loss": 0.5224,
+      "learning_rate": 8.081523993915088e-07,
+      "loss": 0.699,
       "step": 28000
     },
     {
       "epoch": 3.87,
-      "eval_accuracy": 0.7099785630316022,
-      "eval_loss": 1.059259295463562,
-      "eval_runtime": 152.2364,
-      "eval_samples_per_second": 189.981,
-      "eval_steps_per_second": 23.753,
+      "eval_accuracy": 0.7164788050618907,
+      "eval_loss": 0.984923243522644,
+      "eval_runtime": 156.4441,
+      "eval_samples_per_second": 184.871,
+      "eval_steps_per_second": 23.114,
       "step": 28000
     },
     {
       "epoch": 4.0,
       "step": 28924,
       "total_flos": 7.174773947150918e+19,
-      "train_loss": 0.8206716175314105,
-      "train_runtime": 13827.0136,
-      "train_samples_per_second": 66.935,
-      "train_steps_per_second": 2.092
+      "train_loss": 0.9355616505553886,
+      "train_runtime": 13547.1303,
+      "train_samples_per_second": 68.318,
+      "train_steps_per_second": 2.135
     }
   ],
   "max_steps": 28924,