Training in progress, step 100

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 6.0,
-    "eval_accuracy": 0.9924812030075187,
-    "eval_loss": 0.332000732421875,
+    "eval_accuracy": 0.9774436090225563,
+    "eval_loss": 0.195668563246727,
     "eval_model_preparation_time": 0.0027,
-    "eval_runtime": 4.4666,
-    "eval_samples_per_second": 29.777,
-    "eval_steps_per_second": 3.806,
+    "eval_runtime": 5.1924,
+    "eval_samples_per_second": 25.614,
+    "eval_steps_per_second": 3.274,
     "total_flos": 1.542108645707858e+17,
-    "train_loss": 0.5066085149080325,
-    "train_runtime": 701.1808,
-    "train_samples_per_second": 8.848,
-    "train_steps_per_second": 0.556
+    "train_loss": 0.31401297006851586,
+    "train_runtime": 752.1727,
+    "train_samples_per_second": 8.248,
+    "train_steps_per_second": 0.518
 }

config.json

@@ -47,6 +47,7 @@
   ],
   "patch_size": 4,
   "path_norm": true,
+  "problem_type": "single_label_classification",
   "qkv_bias": true,
   "stage_names": [
     "stem",

eval_results.json

@@ -1,9 +1,8 @@
 {
     "epoch": 6.0,
-    "eval_accuracy": 0.9924812030075187,
-    "eval_loss": 0.332000732421875,
-    "eval_model_preparation_time": 0.0027,
-    "eval_runtime": 4.4666,
-    "eval_samples_per_second": 29.777,
-    "eval_steps_per_second": 3.806
+    "eval_accuracy": 0.9774436090225563,
+    "eval_loss": 0.195668563246727,
+    "eval_runtime": 5.1924,
+    "eval_samples_per_second": 25.614,
+    "eval_steps_per_second": 3.274
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e1a1908437a8dd010befaf88307392e437888fd8f76f12b4b3ab777ad71afee
+oid sha256:878abe600651b9c583e4a99ab13b75290089f62371fafcb23dfc40272c67c1cf
 size 110345908

runs/Dec03_16-54-59_l-hvjvvxgp4c.local/events.out.tfevents.1733263670.l-hvjvvxgp4c.local.90109.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa149b44297585d24cf6d3bdea6676805b4c9892e435e6d50ed0c0e93ff11d90
+size 411

runs/Dec03_17-10-01_l-hvjvvxgp4c.local/events.out.tfevents.1733263803.l-hvjvvxgp4c.local.90109.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:315b38b6e47cc4a1af476e1e771f339427adc91dbfca5cb1fc0ea9e9a7693eaa
+size 7849

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 6.0,
     "total_flos": 1.542108645707858e+17,
-    "train_loss": 0.5066085149080325,
-    "train_runtime": 701.1808,
-    "train_samples_per_second": 8.848,
-    "train_steps_per_second": 0.556
+    "train_loss": 0.31401297006851586,
+    "train_runtime": 752.1727,
+    "train_samples_per_second": 8.248,
+    "train_steps_per_second": 0.518
 }

trainer_state.json

@@ -1,331 +1,131 @@
 {
-  "best_metric": 0.34840309619903564,
-  "best_model_checkpoint": "plant_classification/checkpoint-300",
+  "best_metric": 0.1800404042005539,
+  "best_model_checkpoint": "plant_classification/checkpoint-260",
   "epoch": 6.0,
-  "eval_steps": 100,
+  "eval_steps": 500,
   "global_step": 390,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
-    {
-      "epoch": 0.15384615384615385,
-      "grad_norm": 6.1840081214904785,
-      "learning_rate": 2.564102564102564e-06,
-      "loss": 1.097,
-      "step": 10
-    },
-    {
-      "epoch": 0.3076923076923077,
-      "grad_norm": 5.466547966003418,
-      "learning_rate": 5.128205128205128e-06,
-      "loss": 1.0789,
-      "step": 20
-    },
-    {
-      "epoch": 0.46153846153846156,
-      "grad_norm": 4.893115997314453,
-      "learning_rate": 7.692307692307694e-06,
-      "loss": 1.0142,
-      "step": 30
-    },
-    {
-      "epoch": 0.6153846153846154,
-      "grad_norm": 6.236436367034912,
-      "learning_rate": 9.971509971509972e-06,
-      "loss": 0.9414,
-      "step": 40
-    },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 5.752598762512207,
-      "learning_rate": 9.686609686609687e-06,
-      "loss": 0.8319,
+      "grad_norm": 9.507234573364258,
+      "learning_rate": 4.594594594594595e-05,
+      "loss": 0.6848,
       "step": 50
     },
     {
-      "epoch": 0.9230769230769231,
-      "grad_norm": 8.083925247192383,
-      "learning_rate": 9.401709401709402e-06,
-      "loss": 0.7291,
-      "step": 60
-    },
-    {
-      "epoch": 1.0769230769230769,
-      "grad_norm": 8.074342727661133,
-      "learning_rate": 9.116809116809118e-06,
-      "loss": 0.6521,
-      "step": 70
-    },
-    {
-      "epoch": 1.2307692307692308,
-      "grad_norm": 5.033174514770508,
-      "learning_rate": 8.831908831908833e-06,
-      "loss": 0.596,
-      "step": 80
-    },
-    {
-      "epoch": 1.3846153846153846,
-      "grad_norm": 8.695470809936523,
-      "learning_rate": 8.547008547008548e-06,
-      "loss": 0.5473,
-      "step": 90
-    },
-    {
-      "epoch": 1.5384615384615383,
-      "grad_norm": 8.49801254272461,
-      "learning_rate": 8.262108262108262e-06,
-      "loss": 0.484,
-      "step": 100
+      "epoch": 1.0,
+      "eval_accuracy": 0.9699248120300752,
+      "eval_loss": 0.23802712559700012,
+      "eval_runtime": 5.5993,
+      "eval_samples_per_second": 23.753,
+      "eval_steps_per_second": 3.036,
+      "step": 65
     },
     {
       "epoch": 1.5384615384615383,
-      "eval_accuracy": 0.9097744360902256,
-      "eval_loss": 0.4657277762889862,
-      "eval_model_preparation_time": 0.0027,
-      "eval_runtime": 5.3487,
-      "eval_samples_per_second": 24.866,
-      "eval_steps_per_second": 3.178,
+      "grad_norm": 6.534758567810059,
+      "learning_rate": 3.918918918918919e-05,
+      "loss": 0.3222,
       "step": 100
     },
-    {
-      "epoch": 1.6923076923076923,
-      "grad_norm": 5.612356662750244,
-      "learning_rate": 7.977207977207979e-06,
-      "loss": 0.4621,
-      "step": 110
-    },
-    {
-      "epoch": 1.8461538461538463,
-      "grad_norm": 3.638047933578491,
-      "learning_rate": 7.692307692307694e-06,
-      "loss": 0.4771,
-      "step": 120
-    },
     {
       "epoch": 2.0,
-      "grad_norm": 10.50847339630127,
-      "learning_rate": 7.4074074074074075e-06,
-      "loss": 0.452,
+      "eval_accuracy": 0.9849624060150376,
+      "eval_loss": 0.2236112356185913,
+      "eval_runtime": 7.055,
+      "eval_samples_per_second": 18.852,
+      "eval_steps_per_second": 2.41,
       "step": 130
     },
-    {
-      "epoch": 2.1538461538461537,
-      "grad_norm": 5.7228593826293945,
-      "learning_rate": 7.122507122507123e-06,
-      "loss": 0.4428,
-      "step": 140
-    },
     {
       "epoch": 2.3076923076923075,
-      "grad_norm": 7.263442516326904,
-      "learning_rate": 6.837606837606839e-06,
-      "loss": 0.4217,
+      "grad_norm": 5.372588157653809,
+      "learning_rate": 3.2432432432432436e-05,
+      "loss": 0.2839,
       "step": 150
     },
     {
-      "epoch": 2.4615384615384617,
-      "grad_norm": 7.28318452835083,
-      "learning_rate": 6.552706552706553e-06,
-      "loss": 0.4084,
-      "step": 160
-    },
-    {
-      "epoch": 2.6153846153846154,
-      "grad_norm": 8.621001243591309,
-      "learning_rate": 6.267806267806268e-06,
-      "loss": 0.4402,
-      "step": 170
-    },
-    {
-      "epoch": 2.769230769230769,
-      "grad_norm": 9.766897201538086,
-      "learning_rate": 5.982905982905983e-06,
-      "loss": 0.4183,
-      "step": 180
-    },
-    {
-      "epoch": 2.9230769230769234,
-      "grad_norm": 5.3531975746154785,
-      "learning_rate": 5.6980056980056985e-06,
-      "loss": 0.4135,
-      "step": 190
-    },
-    {
-      "epoch": 3.076923076923077,
-      "grad_norm": 5.893077373504639,
-      "learning_rate": 5.413105413105414e-06,
-      "loss": 0.4002,
-      "step": 200
+      "epoch": 3.0,
+      "eval_accuracy": 0.9924812030075187,
+      "eval_loss": 0.20667874813079834,
+      "eval_runtime": 6.7802,
+      "eval_samples_per_second": 19.616,
+      "eval_steps_per_second": 2.507,
+      "step": 195
     },
     {
       "epoch": 3.076923076923077,
-      "eval_accuracy": 0.9699248120300752,
-      "eval_loss": 0.3610040843486786,
-      "eval_model_preparation_time": 0.0027,
-      "eval_runtime": 5.6532,
-      "eval_samples_per_second": 23.527,
-      "eval_steps_per_second": 3.007,
+      "grad_norm": 4.6384596824646,
+      "learning_rate": 2.5675675675675675e-05,
+      "loss": 0.2715,
       "step": 200
     },
-    {
-      "epoch": 3.230769230769231,
-      "grad_norm": 5.7207183837890625,
-      "learning_rate": 5.128205128205128e-06,
-      "loss": 0.3907,
-      "step": 210
-    },
-    {
-      "epoch": 3.3846153846153846,
-      "grad_norm": 5.367263317108154,
-      "learning_rate": 4.8433048433048435e-06,
-      "loss": 0.4029,
-      "step": 220
-    },
-    {
-      "epoch": 3.5384615384615383,
-      "grad_norm": 9.18671989440918,
-      "learning_rate": 4.558404558404559e-06,
-      "loss": 0.4166,
-      "step": 230
-    },
-    {
-      "epoch": 3.6923076923076925,
-      "grad_norm": 5.503479480743408,
-      "learning_rate": 4.273504273504274e-06,
-      "loss": 0.4059,
-      "step": 240
-    },
     {
       "epoch": 3.8461538461538463,
-      "grad_norm": 5.8476786613464355,
-      "learning_rate": 3.9886039886039894e-06,
-      "loss": 0.3642,
+      "grad_norm": 11.217158317565918,
+      "learning_rate": 1.891891891891892e-05,
+      "loss": 0.2422,
       "step": 250
     },
     {
       "epoch": 4.0,
-      "grad_norm": 7.110011577606201,
-      "learning_rate": 3.7037037037037037e-06,
-      "loss": 0.4197,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.1800404042005539,
+      "eval_runtime": 6.8457,
+      "eval_samples_per_second": 19.428,
+      "eval_steps_per_second": 2.483,
       "step": 260
     },
-    {
-      "epoch": 4.153846153846154,
-      "grad_norm": 9.354405403137207,
-      "learning_rate": 3.4188034188034193e-06,
-      "loss": 0.3733,
-      "step": 270
-    },
-    {
-      "epoch": 4.3076923076923075,
-      "grad_norm": 3.5231635570526123,
-      "learning_rate": 3.133903133903134e-06,
-      "loss": 0.4287,
-      "step": 280
-    },
-    {
-      "epoch": 4.461538461538462,
-      "grad_norm": 4.517345905303955,
-      "learning_rate": 2.8490028490028492e-06,
-      "loss": 0.3868,
-      "step": 290
-    },
     {
       "epoch": 4.615384615384615,
-      "grad_norm": 10.131566047668457,
-      "learning_rate": 2.564102564102564e-06,
-      "loss": 0.4059,
+      "grad_norm": 8.160284042358398,
+      "learning_rate": 1.2162162162162164e-05,
+      "loss": 0.246,
       "step": 300
     },
     {
-      "epoch": 4.615384615384615,
-      "eval_accuracy": 0.9849624060150376,
-      "eval_loss": 0.34840309619903564,
-      "eval_model_preparation_time": 0.0027,
-      "eval_runtime": 6.7969,
-      "eval_samples_per_second": 19.568,
-      "eval_steps_per_second": 2.501,
-      "step": 300
-    },
-    {
-      "epoch": 4.769230769230769,
-      "grad_norm": 3.7114639282226562,
-      "learning_rate": 2.2792022792022796e-06,
-      "loss": 0.4177,
-      "step": 310
-    },
-    {
-      "epoch": 4.923076923076923,
-      "grad_norm": 5.1394124031066895,
-      "learning_rate": 1.9943019943019947e-06,
-      "loss": 0.3838,
-      "step": 320
-    },
-    {
-      "epoch": 5.076923076923077,
-      "grad_norm": 5.99159049987793,
-      "learning_rate": 1.7094017094017097e-06,
-      "loss": 0.348,
-      "step": 330
-    },
-    {
-      "epoch": 5.230769230769231,
-      "grad_norm": 2.067537307739258,
-      "learning_rate": 1.4245014245014246e-06,
-      "loss": 0.3965,
-      "step": 340
+      "epoch": 5.0,
+      "eval_accuracy": 0.9774436090225563,
+      "eval_loss": 0.1996687650680542,
+      "eval_runtime": 6.8876,
+      "eval_samples_per_second": 19.31,
+      "eval_steps_per_second": 2.468,
+      "step": 325
     },
     {
       "epoch": 5.384615384615385,
-      "grad_norm": 8.420255661010742,
-      "learning_rate": 1.1396011396011398e-06,
-      "loss": 0.403,
+      "grad_norm": 4.983973979949951,
+      "learning_rate": 5.405405405405406e-06,
+      "loss": 0.2151,
       "step": 350
     },
-    {
-      "epoch": 5.538461538461538,
-      "grad_norm": 5.098587512969971,
-      "learning_rate": 8.547008547008548e-07,
-      "loss": 0.3657,
-      "step": 360
-    },
-    {
-      "epoch": 5.6923076923076925,
-      "grad_norm": 7.574987411499023,
-      "learning_rate": 5.698005698005699e-07,
-      "loss": 0.3783,
-      "step": 370
-    },
-    {
-      "epoch": 5.846153846153846,
-      "grad_norm": 8.597725868225098,
-      "learning_rate": 2.8490028490028494e-07,
-      "loss": 0.3845,
-      "step": 380
-    },
     {
       "epoch": 6.0,
-      "grad_norm": 16.919288635253906,
-      "learning_rate": 0.0,
-      "loss": 0.3776,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.18075048923492432,
+      "eval_runtime": 7.7467,
+      "eval_samples_per_second": 17.169,
+      "eval_steps_per_second": 2.194,
       "step": 390
     },
     {
       "epoch": 6.0,
       "step": 390,
       "total_flos": 1.542108645707858e+17,
-      "train_loss": 0.5066085149080325,
-      "train_runtime": 701.1808,
-      "train_samples_per_second": 8.848,
-      "train_steps_per_second": 0.556
+      "train_loss": 0.31401297006851586,
+      "train_runtime": 752.1727,
+      "train_samples_per_second": 8.248,
+      "train_steps_per_second": 0.518
     }
   ],
-  "logging_steps": 10,
+  "logging_steps": 50,
   "max_steps": 390,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 6,
-  "save_steps": 100,
+  "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
       "args": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b6dac486ee10a56bc8293ae2538c014911af98251123a5cc54d47afbcea2aeaa
+oid sha256:5ed5b4e965288739cb270d4b700f8afb5484490252ad7bc6b6eaae9dc38805dc
 size 5304