Training in progress, epoch 1

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.1,
-    "eval_accuracy": 0.7073170731707317,
-    "eval_loss": 1.3381569385528564,
-    "eval_runtime": 5.7056,
-    "eval_samples_per_second": 14.372,
-    "eval_steps_per_second": 3.681
+    "eval_accuracy": 0.6951219512195121,
+    "eval_loss": 1.2568752765655518,
+    "eval_runtime": 5.6457,
+    "eval_samples_per_second": 14.524,
+    "eval_steps_per_second": 3.72
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:07380e681d431f424eed6cfd1df25fd4a09f163e41c9122585f79cf3c0fb0bde
+oid sha256:8a99ea9829e1200d06ae401d3bf8e36295f08492aa1c756f8551170726883efa
 size 344952716

runs/Aug10_21-42-01_prod3/events.out.tfevents.1754835019.prod3.218391.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18585cb323fa152689aa6129ba1bcbced696133a5732cab2eb1ab767df83168a
+size 411

runs/Aug11_12-54-44_prod3/events.out.tfevents.1754888092.prod3.683740.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae7b3495745bbb1dc24c763deeed8fe96b91e8e61e5846610e507c01212f2b93
+size 8376

test_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.1,
-    "eval_accuracy": 0.7073170731707317,
-    "eval_loss": 1.3381569385528564,
-    "eval_runtime": 5.7056,
-    "eval_samples_per_second": 14.372,
-    "eval_steps_per_second": 3.681
+    "eval_accuracy": 0.6951219512195121,
+    "eval_loss": 1.2568752765655518,
+    "eval_runtime": 5.6457,
+    "eval_samples_per_second": 14.524,
+    "eval_steps_per_second": 3.72
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_global_step": 575,
-  "best_metric": 0.7073170731707317,
-  "best_model_checkpoint": "./Models/Matcha_clips_224_fintuned_4/checkpoint-575",
+  "best_global_step": 460,
+  "best_metric": 0.6951219512195121,
+  "best_model_checkpoint": "./Models/Matcha_clips_224_fintuned_4/checkpoint-460",
   "epoch": 9.1,
   "eval_steps": 500,
   "global_step": 1150,
@@ -11,915 +11,915 @@
   "log_history": [
     {
       "epoch": 0.008695652173913044,
-      "grad_norm": 6.172690391540527,
+      "grad_norm": 6.4240241050720215,
       "learning_rate": 3.91304347826087e-06,
-      "loss": 1.9556,
+      "loss": 1.9844,
       "step": 10
     },
     {
       "epoch": 0.017391304347826087,
-      "grad_norm": 5.285815238952637,
+      "grad_norm": 5.0679426193237305,
       "learning_rate": 8.26086956521739e-06,
-      "loss": 1.9446,
+      "loss": 1.9772,
       "step": 20
     },
     {
       "epoch": 0.02608695652173913,
-      "grad_norm": 7.100526809692383,
+      "grad_norm": 7.605220794677734,
       "learning_rate": 1.2608695652173914e-05,
-      "loss": 1.8525,
+      "loss": 1.854,
       "step": 30
     },
     {
       "epoch": 0.034782608695652174,
-      "grad_norm": 7.057590961456299,
+      "grad_norm": 8.32487964630127,
       "learning_rate": 1.6956521739130433e-05,
-      "loss": 1.8414,
+      "loss": 1.9078,
       "step": 40
     },
     {
       "epoch": 0.043478260869565216,
-      "grad_norm": 6.360028266906738,
+      "grad_norm": 9.476813316345215,
       "learning_rate": 2.1304347826086958e-05,
-      "loss": 1.6835,
+      "loss": 1.6298,
       "step": 50
     },
     {
       "epoch": 0.05217391304347826,
-      "grad_norm": 6.762235164642334,
+      "grad_norm": 12.970176696777344,
       "learning_rate": 2.5652173913043483e-05,
-      "loss": 1.8069,
+      "loss": 1.791,
       "step": 60
     },
     {
       "epoch": 0.06086956521739131,
-      "grad_norm": 9.110674858093262,
+      "grad_norm": 12.027438163757324,
       "learning_rate": 3e-05,
-      "loss": 1.6742,
+      "loss": 1.6904,
       "step": 70
     },
     {
       "epoch": 0.06956521739130435,
-      "grad_norm": 5.7140631675720215,
+      "grad_norm": 9.076032638549805,
       "learning_rate": 3.4347826086956526e-05,
-      "loss": 1.675,
+      "loss": 1.6026,
       "step": 80
     },
     {
       "epoch": 0.0782608695652174,
-      "grad_norm": 7.441013336181641,
+      "grad_norm": 9.857560157775879,
       "learning_rate": 3.869565217391305e-05,
-      "loss": 1.7402,
+      "loss": 1.8465,
       "step": 90
     },
     {
       "epoch": 0.08695652173913043,
-      "grad_norm": 8.591211318969727,
+      "grad_norm": 10.760807991027832,
       "learning_rate": 4.304347826086957e-05,
-      "loss": 1.6256,
+      "loss": 1.5482,
       "step": 100
     },
     {
       "epoch": 0.09565217391304348,
-      "grad_norm": 11.78553581237793,
+      "grad_norm": 13.033352851867676,
       "learning_rate": 4.739130434782609e-05,
-      "loss": 1.4198,
+      "loss": 1.3814,
       "step": 110
     },
     {
       "epoch": 0.1,
-      "eval_accuracy": 0.4634146341463415,
-      "eval_loss": 1.4963948726654053,
-      "eval_runtime": 6.2391,
-      "eval_samples_per_second": 13.143,
-      "eval_steps_per_second": 3.366,
+      "eval_accuracy": 0.4268292682926829,
+      "eval_loss": 1.5108782052993774,
+      "eval_runtime": 6.2348,
+      "eval_samples_per_second": 13.152,
+      "eval_steps_per_second": 3.368,
       "step": 115
     },
     {
       "epoch": 1.0043478260869565,
-      "grad_norm": 9.049060821533203,
+      "grad_norm": 10.566363334655762,
       "learning_rate": 4.980676328502415e-05,
-      "loss": 1.3079,
+      "loss": 1.2744,
       "step": 120
     },
     {
       "epoch": 1.0130434782608695,
-      "grad_norm": 10.772295951843262,
+      "grad_norm": 9.955045700073242,
       "learning_rate": 4.932367149758454e-05,
-      "loss": 1.0031,
+      "loss": 1.0512,
       "step": 130
     },
     {
       "epoch": 1.0217391304347827,
-      "grad_norm": 20.79416275024414,
+      "grad_norm": 21.17795181274414,
       "learning_rate": 4.884057971014493e-05,
-      "loss": 0.9924,
+      "loss": 1.0042,
       "step": 140
     },
     {
       "epoch": 1.0304347826086957,
-      "grad_norm": 6.840089321136475,
+      "grad_norm": 8.381865501403809,
       "learning_rate": 4.8357487922705316e-05,
-      "loss": 0.7422,
+      "loss": 0.7597,
       "step": 150
     },
     {
       "epoch": 1.0391304347826087,
-      "grad_norm": 9.69269847869873,
+      "grad_norm": 10.130925178527832,
       "learning_rate": 4.7874396135265706e-05,
-      "loss": 1.0388,
+      "loss": 0.9822,
       "step": 160
     },
     {
       "epoch": 1.0478260869565217,
-      "grad_norm": 9.885075569152832,
+      "grad_norm": 17.177282333374023,
       "learning_rate": 4.739130434782609e-05,
-      "loss": 0.9266,
+      "loss": 0.8898,
       "step": 170
     },
     {
       "epoch": 1.0565217391304347,
-      "grad_norm": 16.479467391967773,
+      "grad_norm": 20.3695068359375,
       "learning_rate": 4.690821256038648e-05,
-      "loss": 0.8315,
+      "loss": 0.8532,
       "step": 180
     },
     {
       "epoch": 1.065217391304348,
-      "grad_norm": 22.804908752441406,
+      "grad_norm": 15.664092063903809,
       "learning_rate": 4.642512077294686e-05,
-      "loss": 0.736,
+      "loss": 0.873,
       "step": 190
     },
     {
       "epoch": 1.0739130434782609,
-      "grad_norm": 12.778817176818848,
+      "grad_norm": 16.01185417175293,
       "learning_rate": 4.594202898550725e-05,
-      "loss": 0.8893,
+      "loss": 1.0756,
       "step": 200
     },
     {
       "epoch": 1.0826086956521739,
-      "grad_norm": 18.733671188354492,
+      "grad_norm": 19.479690551757812,
       "learning_rate": 4.545893719806764e-05,
-      "loss": 1.3588,
+      "loss": 1.3473,
       "step": 210
     },
     {
       "epoch": 1.0913043478260869,
-      "grad_norm": 19.088119506835938,
+      "grad_norm": 14.130785942077637,
       "learning_rate": 4.497584541062802e-05,
-      "loss": 0.7065,
+      "loss": 0.7922,
       "step": 220
     },
     {
       "epoch": 1.1,
-      "grad_norm": 10.45570182800293,
+      "grad_norm": 9.247044563293457,
       "learning_rate": 4.449275362318841e-05,
-      "loss": 0.6674,
+      "loss": 0.776,
       "step": 230
     },
     {
       "epoch": 1.1,
-      "eval_accuracy": 0.6585365853658537,
-      "eval_loss": 1.0704329013824463,
-      "eval_runtime": 6.1407,
-      "eval_samples_per_second": 13.354,
-      "eval_steps_per_second": 3.42,
+      "eval_accuracy": 0.573170731707317,
+      "eval_loss": 1.0662370920181274,
+      "eval_runtime": 6.0024,
+      "eval_samples_per_second": 13.661,
+      "eval_steps_per_second": 3.499,
       "step": 230
     },
     {
       "epoch": 2.008695652173913,
-      "grad_norm": 7.579151153564453,
+      "grad_norm": 4.517851829528809,
       "learning_rate": 4.4009661835748794e-05,
-      "loss": 0.2858,
+      "loss": 0.2886,
       "step": 240
     },
     {
       "epoch": 2.017391304347826,
-      "grad_norm": 9.255645751953125,
+      "grad_norm": 3.206162452697754,
       "learning_rate": 4.352657004830918e-05,
-      "loss": 0.5303,
+      "loss": 0.4656,
       "step": 250
     },
     {
       "epoch": 2.026086956521739,
-      "grad_norm": 1.6530203819274902,
+      "grad_norm": 1.1080957651138306,
       "learning_rate": 4.304347826086957e-05,
-      "loss": 0.2094,
+      "loss": 0.2634,
       "step": 260
     },
     {
       "epoch": 2.034782608695652,
-      "grad_norm": 13.71268367767334,
+      "grad_norm": 5.800728797912598,
       "learning_rate": 4.256038647342995e-05,
-      "loss": 0.2021,
+      "loss": 0.222,
       "step": 270
     },
     {
       "epoch": 2.0434782608695654,
-      "grad_norm": 2.5825889110565186,
+      "grad_norm": 0.48055946826934814,
       "learning_rate": 4.207729468599034e-05,
-      "loss": 0.1329,
+      "loss": 0.1454,
       "step": 280
     },
     {
       "epoch": 2.0521739130434784,
-      "grad_norm": 1.7412315607070923,
+      "grad_norm": 2.7402403354644775,
       "learning_rate": 4.1594202898550726e-05,
-      "loss": 0.1377,
+      "loss": 0.1433,
       "step": 290
     },
     {
       "epoch": 2.0608695652173914,
-      "grad_norm": 18.217609405517578,
+      "grad_norm": 8.421381950378418,
       "learning_rate": 4.111111111111111e-05,
-      "loss": 0.2466,
+      "loss": 0.3145,
       "step": 300
     },
     {
       "epoch": 2.0695652173913044,
-      "grad_norm": 7.633990287780762,
+      "grad_norm": 4.506998538970947,
       "learning_rate": 4.06280193236715e-05,
-      "loss": 0.3553,
+      "loss": 0.2138,
       "step": 310
     },
     {
       "epoch": 2.0782608695652174,
-      "grad_norm": 19.997325897216797,
+      "grad_norm": 15.715832710266113,
       "learning_rate": 4.014492753623188e-05,
-      "loss": 0.3189,
+      "loss": 0.2296,
       "step": 320
     },
     {
       "epoch": 2.0869565217391304,
-      "grad_norm": 1.3608593940734863,
+      "grad_norm": 1.7134050130844116,
       "learning_rate": 3.966183574879227e-05,
-      "loss": 0.1477,
+      "loss": 0.0711,
       "step": 330
     },
     {
       "epoch": 2.0956521739130434,
-      "grad_norm": 6.2525739669799805,
+      "grad_norm": 3.08317232131958,
       "learning_rate": 3.9178743961352657e-05,
-      "loss": 0.4521,
+      "loss": 0.3286,
       "step": 340
     },
     {
       "epoch": 2.1,
-      "eval_accuracy": 0.6341463414634146,
-      "eval_loss": 0.9644566774368286,
-      "eval_runtime": 6.2182,
-      "eval_samples_per_second": 13.187,
-      "eval_steps_per_second": 3.377,
+      "eval_accuracy": 0.6463414634146342,
+      "eval_loss": 1.0357838869094849,
+      "eval_runtime": 6.0464,
+      "eval_samples_per_second": 13.562,
+      "eval_steps_per_second": 3.473,
       "step": 345
     },
     {
       "epoch": 3.0043478260869567,
-      "grad_norm": 0.2062215656042099,
+      "grad_norm": 0.37449759244918823,
       "learning_rate": 3.869565217391305e-05,
-      "loss": 0.0714,
+      "loss": 0.1835,
       "step": 350
     },
     {
       "epoch": 3.0130434782608697,
-      "grad_norm": 0.07909725606441498,
+      "grad_norm": 0.18086479604244232,
       "learning_rate": 3.821256038647344e-05,
-      "loss": 0.0455,
+      "loss": 0.0584,
       "step": 360
     },
     {
       "epoch": 3.0217391304347827,
-      "grad_norm": 2.4341115951538086,
+      "grad_norm": 0.32367080450057983,
       "learning_rate": 3.772946859903382e-05,
-      "loss": 0.0473,
+      "loss": 0.061,
       "step": 370
     },
     {
       "epoch": 3.0304347826086957,
-      "grad_norm": 0.42377185821533203,
+      "grad_norm": 0.264346718788147,
       "learning_rate": 3.7246376811594204e-05,
-      "loss": 0.111,
+      "loss": 0.0423,
       "step": 380
     },
     {
       "epoch": 3.0391304347826087,
-      "grad_norm": 0.19808898866176605,
+      "grad_norm": 8.53426456451416,
       "learning_rate": 3.6763285024154594e-05,
-      "loss": 0.0683,
+      "loss": 0.0256,
       "step": 390
     },
     {
       "epoch": 3.0478260869565217,
-      "grad_norm": 0.8478032946586609,
+      "grad_norm": 0.227620467543602,
       "learning_rate": 3.628019323671498e-05,
-      "loss": 0.0441,
+      "loss": 0.0892,
       "step": 400
     },
     {
       "epoch": 3.0565217391304347,
-      "grad_norm": 0.15200255811214447,
+      "grad_norm": 0.21275293827056885,
       "learning_rate": 3.579710144927537e-05,
-      "loss": 0.0142,
+      "loss": 0.0628,
       "step": 410
     },
     {
       "epoch": 3.0652173913043477,
-      "grad_norm": 0.1159711629152298,
+      "grad_norm": 0.09173234552145004,
       "learning_rate": 3.531400966183575e-05,
-      "loss": 0.0203,
+      "loss": 0.0088,
       "step": 420
     },
     {
       "epoch": 3.0739130434782607,
-      "grad_norm": 0.1384093165397644,
+      "grad_norm": 0.3767607510089874,
       "learning_rate": 3.4830917874396135e-05,
-      "loss": 0.1054,
+      "loss": 0.0297,
       "step": 430
     },
     {
       "epoch": 3.082608695652174,
-      "grad_norm": 0.5738388895988464,
+      "grad_norm": 0.8464566469192505,
       "learning_rate": 3.4347826086956526e-05,
-      "loss": 0.0707,
+      "loss": 0.0493,
       "step": 440
     },
     {
       "epoch": 3.091304347826087,
-      "grad_norm": 0.16567471623420715,
+      "grad_norm": 0.3951728045940399,
       "learning_rate": 3.386473429951691e-05,
-      "loss": 0.1002,
+      "loss": 0.0953,
       "step": 450
     },
     {
       "epoch": 3.1,
-      "grad_norm": 0.6657196283340454,
+      "grad_norm": 1.0671485662460327,
       "learning_rate": 3.338164251207729e-05,
-      "loss": 0.1047,
+      "loss": 0.1288,
       "step": 460
     },
     {
       "epoch": 3.1,
-      "eval_accuracy": 0.6463414634146342,
-      "eval_loss": 1.2479301691055298,
-      "eval_runtime": 6.3163,
-      "eval_samples_per_second": 12.982,
-      "eval_steps_per_second": 3.325,
+      "eval_accuracy": 0.6951219512195121,
+      "eval_loss": 1.2568752765655518,
+      "eval_runtime": 6.1334,
+      "eval_samples_per_second": 13.369,
+      "eval_steps_per_second": 3.424,
       "step": 460
     },
     {
       "epoch": 4.008695652173913,
-      "grad_norm": 2.4936015605926514,
+      "grad_norm": 3.649686813354492,
       "learning_rate": 3.289855072463768e-05,
-      "loss": 0.0126,
+      "loss": 0.0122,
       "step": 470
     },
     {
       "epoch": 4.017391304347826,
-      "grad_norm": 25.129255294799805,
+      "grad_norm": 1.168199062347412,
       "learning_rate": 3.2415458937198066e-05,
-      "loss": 0.0839,
+      "loss": 0.0274,
       "step": 480
     },
     {
       "epoch": 4.026086956521739,
-      "grad_norm": 0.05726484954357147,
+      "grad_norm": 0.3025151789188385,
       "learning_rate": 3.1932367149758457e-05,
-      "loss": 0.0028,
+      "loss": 0.0088,
       "step": 490
     },
     {
       "epoch": 4.034782608695652,
-      "grad_norm": 0.034710388630628586,
+      "grad_norm": 0.033977147191762924,
       "learning_rate": 3.144927536231884e-05,
-      "loss": 0.0035,
+      "loss": 0.0049,
       "step": 500
     },
     {
       "epoch": 4.043478260869565,
-      "grad_norm": 0.04727836325764656,
+      "grad_norm": 0.22555121779441833,
       "learning_rate": 3.0966183574879224e-05,
-      "loss": 0.0505,
+      "loss": 0.0073,
       "step": 510
     },
     {
       "epoch": 4.052173913043478,
-      "grad_norm": 48.47065353393555,
+      "grad_norm": 1.255247712135315,
       "learning_rate": 3.0483091787439617e-05,
-      "loss": 0.1449,
+      "loss": 0.0074,
       "step": 520
     },
     {
       "epoch": 4.060869565217391,
-      "grad_norm": 0.09721066057682037,
+      "grad_norm": 0.27926549315452576,
       "learning_rate": 3e-05,
-      "loss": 0.0098,
+      "loss": 0.0117,
       "step": 530
     },
     {
       "epoch": 4.069565217391304,
-      "grad_norm": 0.3183492422103882,
+      "grad_norm": 0.019165927544236183,
       "learning_rate": 2.951690821256039e-05,
-      "loss": 0.0041,
+      "loss": 0.002,
       "step": 540
     },
     {
       "epoch": 4.078260869565217,
-      "grad_norm": 0.02137056551873684,
+      "grad_norm": 0.042612988501787186,
       "learning_rate": 2.9033816425120775e-05,
-      "loss": 0.0696,
+      "loss": 0.0192,
       "step": 550
     },
     {
       "epoch": 4.086956521739131,
-      "grad_norm": 0.04266425967216492,
+      "grad_norm": 0.05526461824774742,
       "learning_rate": 2.8550724637681158e-05,
-      "loss": 0.004,
+      "loss": 0.0021,
       "step": 560
     },
     {
       "epoch": 4.095652173913043,
-      "grad_norm": 0.04853034391999245,
+      "grad_norm": 0.03402889147400856,
       "learning_rate": 2.806763285024155e-05,
-      "loss": 0.0315,
+      "loss": 0.0023,
       "step": 570
     },
     {
       "epoch": 4.1,
-      "eval_accuracy": 0.7073170731707317,
-      "eval_loss": 1.338157057762146,
-      "eval_runtime": 6.3184,
-      "eval_samples_per_second": 12.978,
-      "eval_steps_per_second": 3.324,
+      "eval_accuracy": 0.6951219512195121,
+      "eval_loss": 1.3209176063537598,
+      "eval_runtime": 6.1384,
+      "eval_samples_per_second": 13.359,
+      "eval_steps_per_second": 3.421,
       "step": 575
     },
     {
       "epoch": 5.004347826086956,
-      "grad_norm": 0.028169039636850357,
+      "grad_norm": 0.023641662672162056,
       "learning_rate": 2.7584541062801932e-05,
-      "loss": 0.0025,
+      "loss": 0.0017,
       "step": 580
     },
     {
       "epoch": 5.01304347826087,
-      "grad_norm": 0.04628513753414154,
+      "grad_norm": 0.029742104932665825,
       "learning_rate": 2.7101449275362322e-05,
-      "loss": 0.0033,
+      "loss": 0.0014,
       "step": 590
     },
     {
       "epoch": 5.021739130434782,
-      "grad_norm": 0.19218483567237854,
+      "grad_norm": 0.1183994710445404,
       "learning_rate": 2.6618357487922706e-05,
-      "loss": 0.0018,
+      "loss": 0.0016,
       "step": 600
     },
     {
       "epoch": 5.030434782608696,
-      "grad_norm": 0.04920341819524765,
+      "grad_norm": 0.033052194863557816,
       "learning_rate": 2.6135265700483093e-05,
-      "loss": 0.0011,
+      "loss": 0.001,
       "step": 610
     },
     {
       "epoch": 5.039130434782609,
-      "grad_norm": 0.015601912513375282,
+      "grad_norm": 0.016786742955446243,
       "learning_rate": 2.5652173913043483e-05,
-      "loss": 0.0029,
+      "loss": 0.0023,
       "step": 620
     },
     {
       "epoch": 5.047826086956522,
-      "grad_norm": 0.041672300547361374,
+      "grad_norm": 0.020967165008187294,
       "learning_rate": 2.5169082125603866e-05,
-      "loss": 0.0095,
+      "loss": 0.0037,
       "step": 630
     },
     {
       "epoch": 5.056521739130435,
-      "grad_norm": 0.031528741121292114,
+      "grad_norm": 0.07056716829538345,
       "learning_rate": 2.4685990338164253e-05,
-      "loss": 0.0053,
+      "loss": 0.004,
       "step": 640
     },
     {
       "epoch": 5.065217391304348,
-      "grad_norm": 0.06605325639247894,
+      "grad_norm": 0.993996798992157,
       "learning_rate": 2.420289855072464e-05,
-      "loss": 0.0013,
+      "loss": 0.0022,
       "step": 650
     },
     {
       "epoch": 5.073913043478261,
-      "grad_norm": 0.05129764974117279,
+      "grad_norm": 0.028120990842580795,
       "learning_rate": 2.3719806763285024e-05,
-      "loss": 0.0011,
+      "loss": 0.001,
       "step": 660
     },
     {
       "epoch": 5.082608695652174,
-      "grad_norm": 0.06676599383354187,
+      "grad_norm": 0.025797342881560326,
       "learning_rate": 2.323671497584541e-05,
-      "loss": 0.0014,
+      "loss": 0.0013,
       "step": 670
     },
     {
       "epoch": 5.091304347826087,
-      "grad_norm": 0.03804684802889824,
+      "grad_norm": 0.054205186665058136,
       "learning_rate": 2.2753623188405797e-05,
-      "loss": 0.0115,
+      "loss": 0.0045,
       "step": 680
     },
     {
       "epoch": 5.1,
-      "grad_norm": 0.027047328650951385,
+      "grad_norm": 0.02057654596865177,
       "learning_rate": 2.2270531400966184e-05,
       "loss": 0.0012,
       "step": 690
     },
     {
       "epoch": 5.1,
-      "eval_accuracy": 0.6829268292682927,
-      "eval_loss": 1.3492704629898071,
-      "eval_runtime": 6.0789,
-      "eval_samples_per_second": 13.489,
-      "eval_steps_per_second": 3.455,
+      "eval_accuracy": 0.6707317073170732,
+      "eval_loss": 1.2574602365493774,
+      "eval_runtime": 6.097,
+      "eval_samples_per_second": 13.449,
+      "eval_steps_per_second": 3.444,
       "step": 690
     },
     {
       "epoch": 6.008695652173913,
-      "grad_norm": 0.0289983619004488,
+      "grad_norm": 0.028874915093183517,
       "learning_rate": 2.178743961352657e-05,
-      "loss": 0.001,
+      "loss": 0.0009,
       "step": 700
     },
     {
       "epoch": 6.017391304347826,
-      "grad_norm": 0.015507020987570286,
+      "grad_norm": 0.01535830833017826,
       "learning_rate": 2.1304347826086958e-05,
-      "loss": 0.0014,
+      "loss": 0.0011,
       "step": 710
     },
     {
       "epoch": 6.026086956521739,
-      "grad_norm": 0.026568621397018433,
+      "grad_norm": 0.020613593980669975,
       "learning_rate": 2.0821256038647345e-05,
-      "loss": 0.0011,
+      "loss": 0.0028,
       "step": 720
     },
     {
       "epoch": 6.034782608695652,
-      "grad_norm": 0.029902568086981773,
+      "grad_norm": 0.031190281733870506,
       "learning_rate": 2.0338164251207732e-05,
       "loss": 0.0009,
       "step": 730
     },
     {
       "epoch": 6.043478260869565,
-      "grad_norm": 0.01757122576236725,
+      "grad_norm": 0.015504710376262665,
       "learning_rate": 1.985507246376812e-05,
       "loss": 0.0008,
       "step": 740
     },
     {
       "epoch": 6.052173913043478,
-      "grad_norm": 0.02430758625268936,
+      "grad_norm": 0.015204375609755516,
       "learning_rate": 1.9371980676328502e-05,
-      "loss": 0.0041,
+      "loss": 0.0018,
       "step": 750
     },
     {
       "epoch": 6.060869565217391,
-      "grad_norm": 0.013940953649580479,
+      "grad_norm": 0.010091892443597317,
       "learning_rate": 1.888888888888889e-05,
-      "loss": 0.0019,
+      "loss": 0.0011,
       "step": 760
     },
     {
       "epoch": 6.069565217391304,
-      "grad_norm": 0.01289789192378521,
+      "grad_norm": 0.027177872136235237,
       "learning_rate": 1.8405797101449276e-05,
-      "loss": 0.001,
+      "loss": 0.0009,
       "step": 770
     },
     {
       "epoch": 6.078260869565217,
-      "grad_norm": 0.012633569538593292,
+      "grad_norm": 0.011965460143983364,
       "learning_rate": 1.7922705314009663e-05,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 780
     },
     {
       "epoch": 6.086956521739131,
-      "grad_norm": 0.02719777263700962,
+      "grad_norm": 0.014813700690865517,
       "learning_rate": 1.7439613526570046e-05,
-      "loss": 0.0013,
+      "loss": 0.001,
       "step": 790
     },
     {
       "epoch": 6.095652173913043,
-      "grad_norm": 0.01544960681349039,
+      "grad_norm": 0.016053922474384308,
       "learning_rate": 1.6956521739130433e-05,
-      "loss": 0.0009,
+      "loss": 0.0008,
       "step": 800
     },
     {
       "epoch": 6.1,
-      "eval_accuracy": 0.7073170731707317,
-      "eval_loss": 1.3622010946273804,
-      "eval_runtime": 6.1913,
-      "eval_samples_per_second": 13.244,
-      "eval_steps_per_second": 3.392,
+      "eval_accuracy": 0.6829268292682927,
+      "eval_loss": 1.2972846031188965,
+      "eval_runtime": 6.1036,
+      "eval_samples_per_second": 13.435,
+      "eval_steps_per_second": 3.441,
       "step": 805
     },
     {
       "epoch": 7.004347826086956,
-      "grad_norm": 0.014636457897722721,
+      "grad_norm": 0.020354464650154114,
       "learning_rate": 1.6473429951690824e-05,
-      "loss": 0.0022,
+      "loss": 0.0014,
       "step": 810
     },
     {
       "epoch": 7.01304347826087,
-      "grad_norm": 0.011960177682340145,
+      "grad_norm": 0.013575357384979725,
       "learning_rate": 1.599033816425121e-05,
       "loss": 0.0009,
       "step": 820
     },
     {
       "epoch": 7.021739130434782,
-      "grad_norm": 0.025420162826776505,
+      "grad_norm": 0.022664641961455345,
       "learning_rate": 1.5507246376811597e-05,
       "loss": 0.0008,
       "step": 830
     },
     {
       "epoch": 7.030434782608696,
-      "grad_norm": 0.015812881290912628,
+      "grad_norm": 0.026450317353010178,
       "learning_rate": 1.5024154589371981e-05,
-      "loss": 0.0009,
+      "loss": 0.0007,
       "step": 840
     },
     {
       "epoch": 7.039130434782609,
-      "grad_norm": 0.012759621255099773,
+      "grad_norm": 0.013138238340616226,
       "learning_rate": 1.4541062801932368e-05,
       "loss": 0.0008,
       "step": 850
     },
     {
       "epoch": 7.047826086956522,
-      "grad_norm": 0.013558686710894108,
+      "grad_norm": 0.02052425779402256,
       "learning_rate": 1.4057971014492755e-05,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 860
     },
     {
       "epoch": 7.056521739130435,
-      "grad_norm": 0.009738407097756863,
+      "grad_norm": 0.012927724979817867,
       "learning_rate": 1.3574879227053142e-05,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 870
     },
     {
       "epoch": 7.065217391304348,
-      "grad_norm": 0.020090168341994286,
+      "grad_norm": 0.014560637064278126,
       "learning_rate": 1.3091787439613527e-05,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 880
     },
     {
       "epoch": 7.073913043478261,
-      "grad_norm": 0.00872560404241085,
+      "grad_norm": 0.008629231713712215,
       "learning_rate": 1.2608695652173914e-05,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 890
     },
     {
       "epoch": 7.082608695652174,
-      "grad_norm": 0.0459994301199913,
+      "grad_norm": 0.0358402356505394,
       "learning_rate": 1.21256038647343e-05,
-      "loss": 0.0009,
+      "loss": 0.0008,
       "step": 900
     },
     {
       "epoch": 7.091304347826087,
-      "grad_norm": 0.015241548418998718,
+      "grad_norm": 0.014950945042073727,
       "learning_rate": 1.1642512077294687e-05,
       "loss": 0.0007,
       "step": 910
     },
     {
       "epoch": 7.1,
-      "grad_norm": 0.019985787570476532,
+      "grad_norm": 0.01364809088408947,
       "learning_rate": 1.1159420289855073e-05,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 920
     },
     {
       "epoch": 7.1,
-      "eval_accuracy": 0.7073170731707317,
-      "eval_loss": 1.3999933004379272,
-      "eval_runtime": 6.3918,
-      "eval_samples_per_second": 12.829,
-      "eval_steps_per_second": 3.285,
+      "eval_accuracy": 0.6829268292682927,
+      "eval_loss": 1.3180800676345825,
+      "eval_runtime": 6.3025,
+      "eval_samples_per_second": 13.011,
+      "eval_steps_per_second": 3.332,
       "step": 920
     },
     {
       "epoch": 8.008695652173913,
-      "grad_norm": 0.009398338384926319,
+      "grad_norm": 0.008946227841079235,
       "learning_rate": 1.067632850241546e-05,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 930
     },
     {
       "epoch": 8.017391304347827,
-      "grad_norm": 0.026491643860936165,
+      "grad_norm": 0.024497035890817642,
       "learning_rate": 1.0193236714975846e-05,
       "loss": 0.0007,
       "step": 940
     },
     {
       "epoch": 8.02608695652174,
-      "grad_norm": 0.015245695598423481,
+      "grad_norm": 0.01063747052103281,
       "learning_rate": 9.710144927536233e-06,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 950
     },
     {
       "epoch": 8.034782608695652,
-      "grad_norm": 0.014611267484724522,
+      "grad_norm": 0.014025327749550343,
       "learning_rate": 9.227053140096618e-06,
       "loss": 0.0008,
       "step": 960
     },
     {
       "epoch": 8.043478260869565,
-      "grad_norm": 0.023189160972833633,
+      "grad_norm": 0.021205030381679535,
       "learning_rate": 8.743961352657005e-06,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 970
     },
     {
       "epoch": 8.052173913043479,
-      "grad_norm": 0.01567276194691658,
+      "grad_norm": 0.009876050055027008,
       "learning_rate": 8.26086956521739e-06,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 980
     },
     {
       "epoch": 8.060869565217391,
-      "grad_norm": 0.0071034678258001804,
+      "grad_norm": 0.007320360280573368,
       "learning_rate": 7.777777777777777e-06,
       "loss": 0.0006,
       "step": 990
     },
     {
       "epoch": 8.069565217391304,
-      "grad_norm": 0.00896136648952961,
+      "grad_norm": 0.007261498365551233,
       "learning_rate": 7.294685990338164e-06,
       "loss": 0.0006,
       "step": 1000
     },
     {
       "epoch": 8.078260869565218,
-      "grad_norm": 0.012383806519210339,
+      "grad_norm": 0.013149461708962917,
       "learning_rate": 6.811594202898551e-06,
       "loss": 0.0008,
       "step": 1010
     },
     {
       "epoch": 8.08695652173913,
-      "grad_norm": 0.02067715860903263,
+      "grad_norm": 0.013871278613805771,
       "learning_rate": 6.328502415458938e-06,
       "loss": 0.0007,
       "step": 1020
     },
     {
       "epoch": 8.095652173913043,
-      "grad_norm": 0.02547420747578144,
+      "grad_norm": 0.018679574131965637,
       "learning_rate": 5.845410628019324e-06,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 1030
     },
     {
       "epoch": 8.1,
-      "eval_accuracy": 0.7073170731707317,
-      "eval_loss": 1.401489496231079,
-      "eval_runtime": 6.1399,
-      "eval_samples_per_second": 13.355,
-      "eval_steps_per_second": 3.42,
+      "eval_accuracy": 0.6829268292682927,
+      "eval_loss": 1.329393982887268,
+      "eval_runtime": 6.1832,
+      "eval_samples_per_second": 13.262,
+      "eval_steps_per_second": 3.396,
       "step": 1035
     },
     {
       "epoch": 9.004347826086956,
-      "grad_norm": 0.012581247836351395,
+      "grad_norm": 0.011226188391447067,
       "learning_rate": 5.36231884057971e-06,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 1040
     },
     {
       "epoch": 9.013043478260869,
-      "grad_norm": 0.016566790640354156,
+      "grad_norm": 0.017457231879234314,
       "learning_rate": 4.879227053140096e-06,
       "loss": 0.0006,
       "step": 1050
     },
     {
       "epoch": 9.021739130434783,
-      "grad_norm": 0.03303120657801628,
+      "grad_norm": 0.023909136652946472,
       "learning_rate": 4.396135265700483e-06,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 1060
     },
     {
       "epoch": 9.030434782608696,
-      "grad_norm": 0.015636784955859184,
+      "grad_norm": 0.00866635050624609,
       "learning_rate": 3.91304347826087e-06,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 1070
     },
     {
       "epoch": 9.039130434782608,
-      "grad_norm": 0.015506476163864136,
+      "grad_norm": 0.013021158054471016,
       "learning_rate": 3.4299516908212565e-06,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 1080
     },
     {
       "epoch": 9.047826086956523,
-      "grad_norm": 0.01221100240945816,
+      "grad_norm": 0.008441662415862083,
       "learning_rate": 2.9468599033816426e-06,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 1090
     },
     {
       "epoch": 9.056521739130435,
-      "grad_norm": 0.011791340075433254,
+      "grad_norm": 0.012296984903514385,
       "learning_rate": 2.463768115942029e-06,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 1100
     },
     {
       "epoch": 9.065217391304348,
-      "grad_norm": 0.025301584973931313,
+      "grad_norm": 0.025126850232481956,
       "learning_rate": 1.9806763285024155e-06,
-      "loss": 0.0008,
+      "loss": 0.0007,
       "step": 1110
     },
     {
       "epoch": 9.07391304347826,
-      "grad_norm": 0.014775657095015049,
+      "grad_norm": 0.012579759582877159,
       "learning_rate": 1.497584541062802e-06,
       "loss": 0.0007,
       "step": 1120
     },
     {
       "epoch": 9.082608695652175,
-      "grad_norm": 0.025027761235833168,
+      "grad_norm": 0.012673572637140751,
       "learning_rate": 1.0144927536231885e-06,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 1130
     },
     {
       "epoch": 9.091304347826087,
-      "grad_norm": 0.007070675026625395,
+      "grad_norm": 0.007186697795987129,
       "learning_rate": 5.314009661835749e-07,
       "loss": 0.0006,
       "step": 1140
     },
     {
       "epoch": 9.1,
-      "grad_norm": 0.011773703619837761,
+      "grad_norm": 0.013606193475425243,
       "learning_rate": 4.8309178743961356e-08,
-      "loss": 0.0007,
+      "loss": 0.0006,
       "step": 1150
     },
     {
       "epoch": 9.1,
-      "eval_accuracy": 0.7073170731707317,
-      "eval_loss": 1.4007967710494995,
-      "eval_runtime": 6.2852,
-      "eval_samples_per_second": 13.046,
-      "eval_steps_per_second": 3.341,
+      "eval_accuracy": 0.6951219512195121,
+      "eval_loss": 1.332796573638916,
+      "eval_runtime": 6.2824,
+      "eval_samples_per_second": 13.052,
+      "eval_steps_per_second": 3.343,
       "step": 1150
     },
     {
       "epoch": 9.1,
       "step": 1150,
       "total_flos": 5.732152700888678e+18,
-      "train_loss": 0.30212519713110575,
-      "train_runtime": 1217.5297,
-      "train_samples_per_second": 3.778,
-      "train_steps_per_second": 0.945
+      "train_loss": 0.30071158163492445,
+      "train_runtime": 1642.1131,
+      "train_samples_per_second": 2.801,
+      "train_steps_per_second": 0.7
     },
     {
       "epoch": 9.1,
-      "eval_accuracy": 0.7073170731707317,
-      "eval_loss": 1.3381569385528564,
-      "eval_runtime": 5.7056,
-      "eval_samples_per_second": 14.372,
-      "eval_steps_per_second": 3.681,
+      "eval_accuracy": 0.6951219512195121,
+      "eval_loss": 1.2568752765655518,
+      "eval_runtime": 5.6457,
+      "eval_samples_per_second": 14.524,
+      "eval_steps_per_second": 3.72,
       "step": 1150
     }
   ],

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9dcd726e58dbf901b06a4f42284e673960d4b6d9185a8974cbd9733ea2202cde
+oid sha256:ae9e0e9937d4d0c4c0ef16698f0cc325dc3c27e015ba4375f247d8997c61dfd6
 size 5368