End of training

README.md

@@ -3,6 +3,9 @@ library_name: transformers
 license: apache-2.0
 base_model: c14kevincardenas/beit-large-patch16-384-limb
 tags:
+- image-regression
+- human-movement
+- vision
 - generated_from_trainer
 model-index:
 - name: limbxy_ext_crop_hands
@@ -14,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # limbxy_ext_crop_hands
 
-This model is a fine-tuned version of [c14kevincardenas/beit-large-patch16-384-limb](https://huggingface.co/c14kevincardenas/beit-large-patch16-384-limb) on an unknown dataset.
+This model is a fine-tuned version of [c14kevincardenas/beit-large-patch16-384-limb](https://huggingface.co/c14kevincardenas/beit-large-patch16-384-limb) on the c14kevincardenas/beta_caller_284_limbxy_ext_crop_hands dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.0130
 - Rmse: 0.1140

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 20.0,
-    "eval_loss": 0.008117899298667908,
-    "eval_rmse": 0.09012944251298904,
-    "eval_runtime": 5.0777,
-    "eval_samples_per_second": 102.998,
-    "eval_steps_per_second": 1.772,
+    "eval_loss": 0.012980646453797817,
+    "eval_rmse": 0.11399345099925995,
+    "eval_runtime": 4.987,
+    "eval_samples_per_second": 104.873,
+    "eval_steps_per_second": 1.805,
     "total_flos": 0.0,
-    "train_loss": 0.028548606136377823,
-    "train_runtime": 1786.5714,
-    "train_samples_per_second": 33.17,
-    "train_steps_per_second": 0.526
+    "train_loss": 0.03554724605476602,
+    "train_runtime": 1780.6867,
+    "train_samples_per_second": 33.279,
+    "train_steps_per_second": 0.528
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "eval_loss": 0.008117899298667908,
-    "eval_rmse": 0.09012944251298904,
-    "eval_runtime": 5.0777,
-    "eval_samples_per_second": 102.998,
-    "eval_steps_per_second": 1.772
+    "eval_loss": 0.012980646453797817,
+    "eval_rmse": 0.11399345099925995,
+    "eval_runtime": 4.987,
+    "eval_samples_per_second": 104.873,
+    "eval_steps_per_second": 1.805
 }

runs/Oct25_16-03-52_galactica.ad.cirange.net/events.out.tfevents.1729874125.galactica.ad.cirange.net.151270.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c31a8e5ab6a3cee7765ba53a47235eb16c989ba0bb44483f453e8682b9e2c1c1
+size 407

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
     "total_flos": 0.0,
-    "train_loss": 0.028548606136377823,
-    "train_runtime": 1786.5714,
-    "train_samples_per_second": 33.17,
-    "train_steps_per_second": 0.526
+    "train_loss": 0.03554724605476602,
+    "train_runtime": 1780.6867,
+    "train_samples_per_second": 33.279,
+    "train_steps_per_second": 0.528
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.008117899298667908,
+  "best_metric": 0.012980646453797817,
   "best_model_checkpoint": "/mnt/ml_drive/kcardenas/limbxy_ext_crop_hands/checkpoint-940",
   "epoch": 20.0,
   "eval_steps": 500,
@@ -10,451 +10,451 @@
   "log_history": [
     {
       "epoch": 0.5319148936170213,
-      "grad_norm": 828477.25,
+      "grad_norm": 623358.4375,
       "learning_rate": 5e-06,
-      "loss": 0.2561,
+      "loss": 0.2234,
       "step": 25
     },
     {
       "epoch": 1.0,
-      "eval_loss": 0.14001771807670593,
-      "eval_rmse": 0.37419962882995605,
-      "eval_runtime": 5.5814,
-      "eval_samples_per_second": 93.705,
-      "eval_steps_per_second": 1.613,
+      "eval_loss": 0.15381918847560883,
+      "eval_rmse": 0.3922234773635864,
+      "eval_runtime": 5.3503,
+      "eval_samples_per_second": 97.751,
+      "eval_steps_per_second": 1.682,
       "step": 47
     },
     {
       "epoch": 1.0638297872340425,
-      "grad_norm": 38252.140625,
+      "grad_norm": 2976699.0,
       "learning_rate": 1e-05,
-      "loss": 0.1555,
+      "loss": 0.1582,
       "step": 50
     },
     {
       "epoch": 1.5957446808510638,
-      "grad_norm": 5285939.0,
+      "grad_norm": 3106934.25,
       "learning_rate": 1.5e-05,
-      "loss": 0.1544,
+      "loss": 0.1462,
       "step": 75
     },
     {
       "epoch": 2.0,
-      "eval_loss": 0.0731724426150322,
-      "eval_rmse": 0.2704622745513916,
-      "eval_runtime": 5.1268,
-      "eval_samples_per_second": 102.013,
-      "eval_steps_per_second": 1.755,
+      "eval_loss": 0.07616347819566727,
+      "eval_rmse": 0.2758934795856476,
+      "eval_runtime": 4.8381,
+      "eval_samples_per_second": 108.101,
+      "eval_steps_per_second": 1.86,
       "step": 94
     },
     {
       "epoch": 2.127659574468085,
-      "grad_norm": 1347811.375,
+      "grad_norm": 1030784.1875,
       "learning_rate": 2e-05,
-      "loss": 0.1012,
+      "loss": 0.0747,
       "step": 100
     },
     {
       "epoch": 2.6595744680851063,
-      "grad_norm": 2469424.25,
+      "grad_norm": 3752779.75,
       "learning_rate": 2.5e-05,
-      "loss": 0.0404,
+      "loss": 0.0446,
       "step": 125
     },
     {
       "epoch": 3.0,
-      "eval_loss": 0.05305316299200058,
-      "eval_rmse": 0.23046264052391052,
-      "eval_runtime": 4.9002,
-      "eval_samples_per_second": 106.731,
-      "eval_steps_per_second": 1.837,
+      "eval_loss": 0.04878520593047142,
+      "eval_rmse": 0.22092542052268982,
+      "eval_runtime": 4.8409,
+      "eval_samples_per_second": 108.038,
+      "eval_steps_per_second": 1.859,
       "step": 141
     },
     {
       "epoch": 3.1914893617021276,
-      "grad_norm": 3102997.0,
+      "grad_norm": 4766084.5,
       "learning_rate": 3e-05,
-      "loss": 0.0493,
+      "loss": 0.0437,
       "step": 150
     },
     {
       "epoch": 3.723404255319149,
-      "grad_norm": 3376308.0,
+      "grad_norm": 3292715.5,
       "learning_rate": 3.5e-05,
-      "loss": 0.0322,
+      "loss": 0.0436,
       "step": 175
     },
     {
       "epoch": 4.0,
-      "eval_loss": 0.015296007506549358,
-      "eval_rmse": 0.12374413758516312,
-      "eval_runtime": 4.9466,
-      "eval_samples_per_second": 105.73,
-      "eval_steps_per_second": 1.819,
+      "eval_loss": 0.017620466649532318,
+      "eval_rmse": 0.13286800682544708,
+      "eval_runtime": 4.9214,
+      "eval_samples_per_second": 106.271,
+      "eval_steps_per_second": 1.829,
       "step": 188
     },
     {
       "epoch": 4.25531914893617,
-      "grad_norm": 2041911.875,
+      "grad_norm": 2775416.75,
       "learning_rate": 4e-05,
-      "loss": 0.0221,
+      "loss": 0.0285,
       "step": 200
     },
     {
       "epoch": 4.787234042553192,
-      "grad_norm": 752186.125,
+      "grad_norm": 379769.03125,
       "learning_rate": 4.5e-05,
-      "loss": 0.0284,
+      "loss": 0.0326,
       "step": 225
     },
     {
       "epoch": 5.0,
-      "eval_loss": 0.02443227916955948,
-      "eval_rmse": 0.1564129739999771,
-      "eval_runtime": 5.1046,
-      "eval_samples_per_second": 102.456,
-      "eval_steps_per_second": 1.763,
+      "eval_loss": 0.03224412351846695,
+      "eval_rmse": 0.1795557141304016,
+      "eval_runtime": 4.8651,
+      "eval_samples_per_second": 107.5,
+      "eval_steps_per_second": 1.85,
       "step": 235
     },
     {
       "epoch": 5.319148936170213,
-      "grad_norm": 1242507.75,
+      "grad_norm": 1152975.25,
       "learning_rate": 5e-05,
-      "loss": 0.0169,
+      "loss": 0.0314,
       "step": 250
     },
     {
       "epoch": 5.851063829787234,
-      "grad_norm": 1721578.125,
+      "grad_norm": 2960943.0,
       "learning_rate": 4.818840579710145e-05,
-      "loss": 0.0252,
+      "loss": 0.0212,
       "step": 275
     },
     {
       "epoch": 6.0,
-      "eval_loss": 0.02938535064458847,
-      "eval_rmse": 0.17153353989124298,
-      "eval_runtime": 5.0847,
-      "eval_samples_per_second": 102.857,
-      "eval_steps_per_second": 1.77,
+      "eval_loss": 0.01861540600657463,
+      "eval_rmse": 0.13647863268852234,
+      "eval_runtime": 5.0083,
+      "eval_samples_per_second": 104.427,
+      "eval_steps_per_second": 1.797,
       "step": 282
     },
     {
       "epoch": 6.382978723404255,
-      "grad_norm": 1635126.75,
+      "grad_norm": 1633797.25,
       "learning_rate": 4.63768115942029e-05,
-      "loss": 0.0254,
+      "loss": 0.0211,
       "step": 300
     },
     {
       "epoch": 6.914893617021277,
-      "grad_norm": 2097984.0,
+      "grad_norm": 4293457.5,
       "learning_rate": 4.456521739130435e-05,
-      "loss": 0.0123,
+      "loss": 0.0334,
       "step": 325
     },
     {
       "epoch": 7.0,
-      "eval_loss": 0.0171244814991951,
-      "eval_rmse": 0.13094477355480194,
-      "eval_runtime": 5.0335,
-      "eval_samples_per_second": 103.904,
-      "eval_steps_per_second": 1.788,
+      "eval_loss": 0.030206123366951942,
+      "eval_rmse": 0.1739334762096405,
+      "eval_runtime": 4.8758,
+      "eval_samples_per_second": 107.264,
+      "eval_steps_per_second": 1.846,
       "step": 329
     },
     {
       "epoch": 7.446808510638298,
-      "grad_norm": 3019492.25,
+      "grad_norm": 4272689.0,
       "learning_rate": 4.27536231884058e-05,
-      "loss": 0.0148,
+      "loss": 0.027,
       "step": 350
     },
     {
       "epoch": 7.9787234042553195,
-      "grad_norm": 3237462.5,
+      "grad_norm": 1988792.5,
       "learning_rate": 4.094202898550725e-05,
-      "loss": 0.0216,
+      "loss": 0.0247,
       "step": 375
     },
     {
       "epoch": 8.0,
-      "eval_loss": 0.015043811872601509,
-      "eval_rmse": 0.12263169139623642,
-      "eval_runtime": 4.8436,
-      "eval_samples_per_second": 107.977,
-      "eval_steps_per_second": 1.858,
+      "eval_loss": 0.03623066842556,
+      "eval_rmse": 0.19043287634849548,
+      "eval_runtime": 5.0165,
+      "eval_samples_per_second": 104.256,
+      "eval_steps_per_second": 1.794,
       "step": 376
     },
     {
       "epoch": 8.51063829787234,
-      "grad_norm": 2551103.0,
+      "grad_norm": 1632416.25,
       "learning_rate": 3.91304347826087e-05,
-      "loss": 0.0101,
+      "loss": 0.0238,
       "step": 400
     },
     {
       "epoch": 9.0,
-      "eval_loss": 0.024509919807314873,
-      "eval_rmse": 0.15662036836147308,
-      "eval_runtime": 5.014,
-      "eval_samples_per_second": 104.309,
-      "eval_steps_per_second": 1.795,
+      "eval_loss": 0.015264267101883888,
+      "eval_rmse": 0.12361017614603043,
+      "eval_runtime": 4.9733,
+      "eval_samples_per_second": 105.161,
+      "eval_steps_per_second": 1.81,
       "step": 423
     },
     {
       "epoch": 9.042553191489361,
-      "grad_norm": 4249340.5,
+      "grad_norm": 1837231.75,
       "learning_rate": 3.7318840579710144e-05,
-      "loss": 0.0096,
+      "loss": 0.0182,
       "step": 425
     },
     {
       "epoch": 9.574468085106384,
-      "grad_norm": 963757.5,
+      "grad_norm": 3900810.5,
       "learning_rate": 3.5507246376811596e-05,
-      "loss": 0.012,
+      "loss": 0.0342,
       "step": 450
     },
     {
       "epoch": 10.0,
-      "eval_loss": 0.011943971738219261,
-      "eval_rmse": 0.10932554304599762,
-      "eval_runtime": 4.9998,
-      "eval_samples_per_second": 104.604,
-      "eval_steps_per_second": 1.8,
+      "eval_loss": 0.015248410403728485,
+      "eval_rmse": 0.1235644593834877,
+      "eval_runtime": 4.9635,
+      "eval_samples_per_second": 105.368,
+      "eval_steps_per_second": 1.813,
       "step": 470
     },
     {
       "epoch": 10.106382978723405,
-      "grad_norm": 1531059.5,
+      "grad_norm": 2050042.25,
       "learning_rate": 3.369565217391305e-05,
-      "loss": 0.0073,
+      "loss": 0.0186,
       "step": 475
     },
     {
       "epoch": 10.638297872340425,
-      "grad_norm": 1131259.625,
+      "grad_norm": 3730581.0,
       "learning_rate": 3.188405797101449e-05,
-      "loss": 0.0083,
+      "loss": 0.0188,
       "step": 500
     },
     {
       "epoch": 11.0,
-      "eval_loss": 0.014895778149366379,
-      "eval_rmse": 0.12203000485897064,
-      "eval_runtime": 5.144,
-      "eval_samples_per_second": 101.671,
-      "eval_steps_per_second": 1.75,
+      "eval_loss": 0.01616613008081913,
+      "eval_rmse": 0.12722349166870117,
+      "eval_runtime": 4.9524,
+      "eval_samples_per_second": 105.606,
+      "eval_steps_per_second": 1.817,
       "step": 517
     },
     {
       "epoch": 11.170212765957446,
-      "grad_norm": 1930957.75,
+      "grad_norm": 865433.75,
       "learning_rate": 3.0072463768115944e-05,
-      "loss": 0.0071,
+      "loss": 0.0163,
       "step": 525
     },
     {
       "epoch": 11.702127659574469,
-      "grad_norm": 1709068.75,
+      "grad_norm": 658995.6875,
       "learning_rate": 2.826086956521739e-05,
-      "loss": 0.0072,
+      "loss": 0.0165,
       "step": 550
     },
     {
       "epoch": 12.0,
-      "eval_loss": 0.009299115277826786,
-      "eval_rmse": 0.09643761068582535,
-      "eval_runtime": 4.9587,
-      "eval_samples_per_second": 105.471,
-      "eval_steps_per_second": 1.815,
+      "eval_loss": 0.022758642211556435,
+      "eval_rmse": 0.15094806253910065,
+      "eval_runtime": 4.8205,
+      "eval_samples_per_second": 108.496,
+      "eval_steps_per_second": 1.867,
       "step": 564
     },
     {
       "epoch": 12.23404255319149,
-      "grad_norm": 1622884.0,
+      "grad_norm": 2234811.25,
       "learning_rate": 2.6449275362318844e-05,
-      "loss": 0.0089,
+      "loss": 0.0179,
       "step": 575
     },
     {
       "epoch": 12.76595744680851,
-      "grad_norm": 780158.875,
+      "grad_norm": 1162747.5,
       "learning_rate": 2.4637681159420292e-05,
-      "loss": 0.006,
+      "loss": 0.0163,
       "step": 600
     },
     {
       "epoch": 13.0,
-      "eval_loss": 0.00986443180590868,
-      "eval_rmse": 0.09930924326181412,
-      "eval_runtime": 5.0729,
-      "eval_samples_per_second": 103.096,
-      "eval_steps_per_second": 1.774,
+      "eval_loss": 0.014487933367490768,
+      "eval_rmse": 0.12045549601316452,
+      "eval_runtime": 5.0569,
+      "eval_samples_per_second": 103.423,
+      "eval_steps_per_second": 1.78,
       "step": 611
     },
     {
       "epoch": 13.297872340425531,
-      "grad_norm": 995551.3125,
+      "grad_norm": 4238076.0,
       "learning_rate": 2.282608695652174e-05,
-      "loss": 0.0049,
+      "loss": 0.0176,
       "step": 625
     },
     {
       "epoch": 13.829787234042554,
-      "grad_norm": 731185.1875,
+      "grad_norm": 172452.703125,
       "learning_rate": 2.101449275362319e-05,
-      "loss": 0.0044,
+      "loss": 0.0154,
       "step": 650
     },
     {
       "epoch": 14.0,
-      "eval_loss": 0.010285141877830029,
-      "eval_rmse": 0.10145359486341476,
-      "eval_runtime": 5.0416,
-      "eval_samples_per_second": 103.736,
-      "eval_steps_per_second": 1.785,
+      "eval_loss": 0.013961490243673325,
+      "eval_rmse": 0.11822474002838135,
+      "eval_runtime": 5.1475,
+      "eval_samples_per_second": 101.603,
+      "eval_steps_per_second": 1.748,
       "step": 658
     },
     {
       "epoch": 14.361702127659575,
-      "grad_norm": 1295382.375,
+      "grad_norm": 1203963.375,
       "learning_rate": 1.9202898550724637e-05,
-      "loss": 0.0058,
+      "loss": 0.0154,
       "step": 675
     },
     {
       "epoch": 14.893617021276595,
-      "grad_norm": 751297.9375,
+      "grad_norm": 1234547.5,
       "learning_rate": 1.739130434782609e-05,
-      "loss": 0.0041,
+      "loss": 0.0168,
       "step": 700
     },
     {
       "epoch": 15.0,
-      "eval_loss": 0.008721875958144665,
-      "eval_rmse": 0.09339401870965958,
-      "eval_runtime": 5.0079,
-      "eval_samples_per_second": 104.434,
-      "eval_steps_per_second": 1.797,
+      "eval_loss": 0.014746125787496567,
+      "eval_rmse": 0.12149151414632797,
+      "eval_runtime": 4.9322,
+      "eval_samples_per_second": 106.038,
+      "eval_steps_per_second": 1.825,
       "step": 705
     },
     {
       "epoch": 15.425531914893616,
-      "grad_norm": 1133890.875,
+      "grad_norm": 878691.75,
       "learning_rate": 1.5579710144927537e-05,
-      "loss": 0.003,
+      "loss": 0.0154,
       "step": 725
     },
     {
       "epoch": 15.957446808510639,
-      "grad_norm": 1333124.375,
+      "grad_norm": 2425489.25,
       "learning_rate": 1.3768115942028985e-05,
-      "loss": 0.003,
+      "loss": 0.0156,
       "step": 750
     },
     {
       "epoch": 16.0,
-      "eval_loss": 0.012206021696329117,
-      "eval_rmse": 0.11044398695230484,
-      "eval_runtime": 4.7907,
-      "eval_samples_per_second": 109.169,
-      "eval_steps_per_second": 1.879,
+      "eval_loss": 0.014588478021323681,
+      "eval_rmse": 0.12084110826253891,
+      "eval_runtime": 5.2911,
+      "eval_samples_per_second": 98.846,
+      "eval_steps_per_second": 1.701,
       "step": 752
     },
     {
       "epoch": 16.48936170212766,
-      "grad_norm": 1097979.375,
+      "grad_norm": 497963.0,
       "learning_rate": 1.1956521739130435e-05,
-      "loss": 0.0044,
+      "loss": 0.0159,
       "step": 775
     },
     {
       "epoch": 17.0,
-      "eval_loss": 0.009473947808146477,
-      "eval_rmse": 0.09737709909677505,
-      "eval_runtime": 4.979,
-      "eval_samples_per_second": 105.041,
-      "eval_steps_per_second": 1.808,
+      "eval_loss": 0.015692666172981262,
+      "eval_rmse": 0.12533892691135406,
+      "eval_runtime": 4.7367,
+      "eval_samples_per_second": 110.415,
+      "eval_steps_per_second": 1.9,
       "step": 799
     },
     {
       "epoch": 17.02127659574468,
-      "grad_norm": 1345082.75,
+      "grad_norm": 1821127.75,
       "learning_rate": 1.0144927536231885e-05,
-      "loss": 0.0023,
+      "loss": 0.0149,
       "step": 800
     },
     {
       "epoch": 17.5531914893617,
-      "grad_norm": 252615.640625,
+      "grad_norm": 772281.125,
       "learning_rate": 8.333333333333334e-06,
-      "loss": 0.0019,
+      "loss": 0.0145,
       "step": 825
     },
     {
       "epoch": 18.0,
-      "eval_loss": 0.0081509193405509,
-      "eval_rmse": 0.09029744565486908,
-      "eval_runtime": 5.0468,
-      "eval_samples_per_second": 103.629,
-      "eval_steps_per_second": 1.783,
+      "eval_loss": 0.014158854261040688,
+      "eval_rmse": 0.11905129253864288,
+      "eval_runtime": 4.935,
+      "eval_samples_per_second": 105.978,
+      "eval_steps_per_second": 1.824,
       "step": 846
     },
     {
       "epoch": 18.085106382978722,
-      "grad_norm": 301606.09375,
+      "grad_norm": 480387.40625,
       "learning_rate": 6.521739130434783e-06,
-      "loss": 0.0017,
+      "loss": 0.0133,
       "step": 850
     },
     {
       "epoch": 18.617021276595743,
-      "grad_norm": 594859.5625,
+      "grad_norm": 622875.125,
       "learning_rate": 4.710144927536232e-06,
-      "loss": 0.0018,
+      "loss": 0.0133,
       "step": 875
     },
     {
       "epoch": 19.0,
-      "eval_loss": 0.009004783816635609,
-      "eval_rmse": 0.09493986517190933,
-      "eval_runtime": 4.9612,
-      "eval_samples_per_second": 105.419,
-      "eval_steps_per_second": 1.814,
+      "eval_loss": 0.013214284554123878,
+      "eval_rmse": 0.11501393467187881,
+      "eval_runtime": 4.9198,
+      "eval_samples_per_second": 106.304,
+      "eval_steps_per_second": 1.829,
       "step": 893
     },
     {
       "epoch": 19.148936170212767,
-      "grad_norm": 798350.8125,
+      "grad_norm": 335946.59375,
       "learning_rate": 2.898550724637681e-06,
-      "loss": 0.0017,
+      "loss": 0.0133,
       "step": 900
     },
     {
       "epoch": 19.680851063829788,
-      "grad_norm": 109003.640625,
+      "grad_norm": 479721.125,
       "learning_rate": 1.0869565217391306e-06,
-      "loss": 0.0014,
+      "loss": 0.0127,
       "step": 925
     },
     {
       "epoch": 20.0,
-      "eval_loss": 0.008117899298667908,
-      "eval_rmse": 0.09012944251298904,
-      "eval_runtime": 4.9089,
-      "eval_samples_per_second": 106.541,
-      "eval_steps_per_second": 1.833,
+      "eval_loss": 0.012980646453797817,
+      "eval_rmse": 0.11399345099925995,
+      "eval_runtime": 4.9566,
+      "eval_samples_per_second": 105.517,
+      "eval_steps_per_second": 1.816,
       "step": 940
     },
     {
       "epoch": 20.0,
       "step": 940,
       "total_flos": 0.0,
-      "train_loss": 0.028548606136377823,
-      "train_runtime": 1786.5714,
-      "train_samples_per_second": 33.17,
-      "train_steps_per_second": 0.526
+      "train_loss": 0.03554724605476602,
+      "train_runtime": 1780.6867,
+      "train_samples_per_second": 33.279,
+      "train_steps_per_second": 0.528
     }
   ],
   "logging_steps": 25,