heado

README.md

@@ -19,8 +19,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the beans dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1842
-- Accuracy: 0.9531
+- Loss: 0.0205
+- Accuracy: 1.0
 
 ## Model description
 
@@ -52,8 +52,8 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss | Accuracy |
 |:-------------:|:------:|:----:|:---------------:|:--------:|
-| 0.0316        | 1.5385 | 100  | 0.0977          | 0.9774   |
-| 0.0106        | 3.0769 | 200  | 0.1979          | 0.9624   |
+| 0.091         | 1.5385 | 100  | 0.0747          | 0.9774   |
+| 0.036         | 3.0769 | 200  | 0.0205          | 1.0      |
 
 
 ### Framework versions

all_results.json

@@ -1,13 +1,8 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 0.953125,
-    "eval_loss": 0.1842205822467804,
-    "eval_runtime": 1.4634,
-    "eval_samples_per_second": 87.465,
-    "eval_steps_per_second": 10.933,
     "total_flos": 3.205097416476426e+17,
-    "train_loss": 0.02204044573009014,
-    "train_runtime": 91.5843,
-    "train_samples_per_second": 45.161,
-    "train_steps_per_second": 2.839
+    "train_loss": 0.11695209254439061,
+    "train_runtime": 89.4129,
+    "train_samples_per_second": 46.257,
+    "train_steps_per_second": 2.908
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e5443676f9f75c55203d1d8085f08d7be4c158d489cce154442d42d958a547b8
+oid sha256:ef21ae29af6781b332bf4ae417d3e2c6af87f6dfcb14dccb2bb0a78d6e8925f6
 size 343227052

runs/Jul04_07-32-00_03d0faa5d47a/events.out.tfevents.1720078574.03d0faa5d47a.1790.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92648a71a91c9c421018e6e20271a1293de1dea06e3ee31b3f603aecd7e4e9c7
+size 11323

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
     "total_flos": 3.205097416476426e+17,
-    "train_loss": 0.02204044573009014,
-    "train_runtime": 91.5843,
-    "train_samples_per_second": 45.161,
-    "train_steps_per_second": 2.839
+    "train_loss": 0.11695209254439061,
+    "train_runtime": 89.4129,
+    "train_samples_per_second": 46.257,
+    "train_steps_per_second": 2.908
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.09767168015241623,
-  "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-100",
+  "best_metric": 0.020527832210063934,
+  "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-200",
   "epoch": 4.0,
   "eval_steps": 100,
   "global_step": 260,
@@ -10,212 +10,212 @@
   "log_history": [
     {
       "epoch": 0.15384615384615385,
-      "grad_norm": 0.0549064539372921,
+      "grad_norm": 1.3131574392318726,
       "learning_rate": 0.00019230769230769233,
-      "loss": 0.011,
+      "loss": 0.8368,
       "step": 10
     },
     {
       "epoch": 0.3076923076923077,
-      "grad_norm": 0.053657062351703644,
+      "grad_norm": 1.571279525756836,
       "learning_rate": 0.00018461538461538463,
-      "loss": 0.0099,
+      "loss": 0.3474,
       "step": 20
     },
     {
       "epoch": 0.46153846153846156,
-      "grad_norm": 0.04586618021130562,
+      "grad_norm": 3.746014356613159,
       "learning_rate": 0.00017692307692307693,
-      "loss": 0.0085,
+      "loss": 0.2139,
       "step": 30
     },
     {
       "epoch": 0.6153846153846154,
-      "grad_norm": 0.03834086284041405,
+      "grad_norm": 1.4054049253463745,
       "learning_rate": 0.00016923076923076923,
-      "loss": 0.0074,
+      "loss": 0.1799,
       "step": 40
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 0.0455704927444458,
+      "grad_norm": 0.2104187309741974,
       "learning_rate": 0.00016153846153846155,
-      "loss": 0.0081,
+      "loss": 0.1475,
       "step": 50
     },
     {
       "epoch": 0.9230769230769231,
-      "grad_norm": 4.641096591949463,
+      "grad_norm": 4.926904678344727,
       "learning_rate": 0.00015384615384615385,
-      "loss": 0.0547,
+      "loss": 0.2264,
       "step": 60
     },
     {
       "epoch": 1.0769230769230769,
-      "grad_norm": 10.7581205368042,
+      "grad_norm": 3.4138243198394775,
       "learning_rate": 0.00014615384615384615,
-      "loss": 0.028,
+      "loss": 0.186,
       "step": 70
     },
     {
       "epoch": 1.2307692307692308,
-      "grad_norm": 0.04066062346100807,
+      "grad_norm": 0.18232814967632294,
       "learning_rate": 0.00013846153846153847,
-      "loss": 0.0387,
+      "loss": 0.1536,
       "step": 80
     },
     {
       "epoch": 1.3846153846153846,
-      "grad_norm": 0.042889270931482315,
+      "grad_norm": 0.23882938921451569,
       "learning_rate": 0.00013076923076923077,
-      "loss": 0.0808,
+      "loss": 0.0561,
       "step": 90
     },
     {
       "epoch": 1.5384615384615383,
-      "grad_norm": 0.039594896137714386,
+      "grad_norm": 1.603104829788208,
       "learning_rate": 0.0001230769230769231,
-      "loss": 0.0316,
+      "loss": 0.091,
       "step": 100
     },
     {
       "epoch": 1.5384615384615383,
       "eval_accuracy": 0.9774436090225563,
-      "eval_loss": 0.09767168015241623,
-      "eval_runtime": 1.2787,
-      "eval_samples_per_second": 104.015,
-      "eval_steps_per_second": 13.295,
+      "eval_loss": 0.07465264946222305,
+      "eval_runtime": 1.2611,
+      "eval_samples_per_second": 105.465,
+      "eval_steps_per_second": 13.48,
       "step": 100
     },
     {
       "epoch": 1.6923076923076923,
-      "grad_norm": 0.4228585660457611,
+      "grad_norm": 0.11675859242677689,
       "learning_rate": 0.00011538461538461538,
-      "loss": 0.0357,
+      "loss": 0.0952,
       "step": 110
     },
     {
       "epoch": 1.8461538461538463,
-      "grad_norm": 0.03368207439780235,
+      "grad_norm": 0.11038266867399216,
       "learning_rate": 0.0001076923076923077,
-      "loss": 0.0721,
+      "loss": 0.1683,
       "step": 120
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.041465017944574356,
+      "grad_norm": 1.329042673110962,
       "learning_rate": 0.0001,
-      "loss": 0.0286,
+      "loss": 0.0489,
       "step": 130
     },
     {
       "epoch": 2.1538461538461537,
-      "grad_norm": 4.102177619934082,
+      "grad_norm": 0.46014711260795593,
       "learning_rate": 9.230769230769232e-05,
-      "loss": 0.0239,
+      "loss": 0.0307,
       "step": 140
     },
     {
       "epoch": 2.3076923076923075,
-      "grad_norm": 1.726514458656311,
+      "grad_norm": 0.6358932852745056,
       "learning_rate": 8.461538461538461e-05,
-      "loss": 0.0085,
+      "loss": 0.0254,
       "step": 150
     },
     {
       "epoch": 2.4615384615384617,
-      "grad_norm": 0.025889277458190918,
+      "grad_norm": 0.07531006634235382,
       "learning_rate": 7.692307692307693e-05,
-      "loss": 0.0052,
+      "loss": 0.0175,
       "step": 160
     },
     {
       "epoch": 2.6153846153846154,
-      "grad_norm": 0.025063877925276756,
+      "grad_norm": 0.07003136724233627,
       "learning_rate": 6.923076923076924e-05,
-      "loss": 0.0046,
+      "loss": 0.0172,
       "step": 170
     },
     {
       "epoch": 2.769230769230769,
-      "grad_norm": 0.024265997111797333,
+      "grad_norm": 0.08160967379808426,
       "learning_rate": 6.153846153846155e-05,
-      "loss": 0.0046,
+      "loss": 0.0435,
       "step": 180
     },
     {
       "epoch": 2.9230769230769234,
-      "grad_norm": 15.569941520690918,
+      "grad_norm": 0.09463630616664886,
       "learning_rate": 5.384615384615385e-05,
-      "loss": 0.0232,
+      "loss": 0.0181,
       "step": 190
     },
     {
       "epoch": 3.076923076923077,
-      "grad_norm": 0.023741118609905243,
+      "grad_norm": 0.06692061573266983,
       "learning_rate": 4.615384615384616e-05,
-      "loss": 0.0106,
+      "loss": 0.036,
       "step": 200
     },
     {
       "epoch": 3.076923076923077,
-      "eval_accuracy": 0.9624060150375939,
-      "eval_loss": 0.19794000685214996,
-      "eval_runtime": 1.6975,
-      "eval_samples_per_second": 78.351,
-      "eval_steps_per_second": 10.015,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.020527832210063934,
+      "eval_runtime": 1.2254,
+      "eval_samples_per_second": 108.536,
+      "eval_steps_per_second": 13.873,
       "step": 200
     },
     {
       "epoch": 3.230769230769231,
-      "grad_norm": 0.03866928815841675,
+      "grad_norm": 0.057732608169317245,
       "learning_rate": 3.846153846153846e-05,
-      "loss": 0.0377,
+      "loss": 0.0228,
       "step": 210
     },
     {
       "epoch": 3.3846153846153846,
-      "grad_norm": 0.023310711607336998,
+      "grad_norm": 0.06201878562569618,
       "learning_rate": 3.0769230769230774e-05,
-      "loss": 0.0221,
+      "loss": 0.0116,
       "step": 220
     },
     {
       "epoch": 3.5384615384615383,
-      "grad_norm": 0.02209044247865677,
+      "grad_norm": 0.06420481950044632,
       "learning_rate": 2.307692307692308e-05,
-      "loss": 0.0057,
+      "loss": 0.0325,
       "step": 230
     },
     {
       "epoch": 3.6923076923076925,
-      "grad_norm": 0.020965544506907463,
+      "grad_norm": 0.060832858085632324,
       "learning_rate": 1.5384615384615387e-05,
-      "loss": 0.004,
+      "loss": 0.0114,
       "step": 240
     },
     {
       "epoch": 3.8461538461538463,
-      "grad_norm": 0.02124304138123989,
+      "grad_norm": 0.061510924249887466,
       "learning_rate": 7.692307692307694e-06,
-      "loss": 0.0039,
+      "loss": 0.0115,
       "step": 250
     },
     {
       "epoch": 4.0,
-      "grad_norm": 0.0217234306037426,
+      "grad_norm": 0.05591738596558571,
       "learning_rate": 0.0,
-      "loss": 0.0039,
+      "loss": 0.0114,
       "step": 260
     },
     {
       "epoch": 4.0,
       "step": 260,
       "total_flos": 3.205097416476426e+17,
-      "train_loss": 0.02204044573009014,
-      "train_runtime": 91.5843,
-      "train_samples_per_second": 45.161,
-      "train_steps_per_second": 2.839
+      "train_loss": 0.11695209254439061,
+      "train_runtime": 89.4129,
+      "train_samples_per_second": 46.257,
+      "train_steps_per_second": 2.908
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44164b4d4762f13debd0722ab6612948375146c034c87f96580824bc81c4f4c8
+oid sha256:87354a911506538377ec10b629caf2f0beaa9bfa97389a2982938b09900d7a15
 size 5112