Upload trained model

README.md

@@ -21,11 +21,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [BTX24/hiera-base-224-in1k-hf-finetuned-stroke-binary](https://huggingface.co/BTX24/hiera-base-224-in1k-hf-finetuned-stroke-binary) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0628
-- Accuracy: 0.9854
-- F1: 0.9854
-- Precision: 0.9855
-- Recall: 0.9854
+- Loss: 0.0648
+- Accuracy: 0.9861
+- F1: 0.9861
+- Precision: 0.9862
+- Recall: 0.9861
 
 ## Model description
 

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 12.0,
+    "total_flos": 3.1692355393112064e+18,
+    "train_loss": 0.048116779152726954,
+    "train_runtime": 1311.0723,
+    "train_samples_per_second": 52.729,
+    "train_steps_per_second": 0.833
+}

runs/Jun09_15-00-30_236d6e780da0/events.out.tfevents.1749482612.236d6e780da0.523.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b7e8464fec94a78c73988313a1e10bfe2446da2cbefcda02e5a05f0db0b9f77
+size 560

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 12.0,
+    "total_flos": 3.1692355393112064e+18,
+    "train_loss": 0.048116779152726954,
+    "train_runtime": 1311.0723,
+    "train_samples_per_second": 52.729,
+    "train_steps_per_second": 0.833
+}

trainer_state.json

@@ -0,0 +1,310 @@
+{
+  "best_global_step": 800,
+  "best_metric": 0.9861160209367208,
+  "best_model_checkpoint": "hiera-finetuned-stroke-multi/checkpoint-800",
+  "epoch": 12.0,
+  "eval_steps": 100,
+  "global_step": 1092,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.554016620498615,
+      "grad_norm": 12.229708671569824,
+      "learning_rate": 8.727272727272728e-06,
+      "loss": 0.0578,
+      "step": 50
+    },
+    {
+      "epoch": 1.0997229916897506,
+      "grad_norm": Infinity,
+      "learning_rate": 1.781818181818182e-05,
+      "loss": 0.0612,
+      "step": 100
+    },
+    {
+      "epoch": 1.0997229916897506,
+      "eval_accuracy": 0.963913948646773,
+      "eval_f1": 0.9634566601928356,
+      "eval_loss": 0.1258896440267563,
+      "eval_precision": 0.9646840164123915,
+      "eval_recall": 0.963913948646773,
+      "eval_runtime": 11.8652,
+      "eval_samples_per_second": 121.448,
+      "eval_steps_per_second": 15.255,
+      "step": 100
+    },
+    {
+      "epoch": 1.6537396121883656,
+      "grad_norm": 14.079134941101074,
+      "learning_rate": 1.9930024972933966e-05,
+      "loss": 0.0901,
+      "step": 150
+    },
+    {
+      "epoch": 2.1994459833795013,
+      "grad_norm": 6.2245001792907715,
+      "learning_rate": 1.9615160313232122e-05,
+      "loss": 0.0694,
+      "step": 200
+    },
+    {
+      "epoch": 2.1994459833795013,
+      "eval_accuracy": 0.9777931991672449,
+      "eval_f1": 0.9776500351660441,
+      "eval_loss": 0.0734671950340271,
+      "eval_precision": 0.9779683888397975,
+      "eval_recall": 0.9777931991672449,
+      "eval_runtime": 11.5615,
+      "eval_samples_per_second": 124.637,
+      "eval_steps_per_second": 15.655,
+      "step": 200
+    },
+    {
+      "epoch": 2.7534626038781163,
+      "grad_norm": 8.907445907592773,
+      "learning_rate": 1.905479814346962e-05,
+      "loss": 0.0759,
+      "step": 250
+    },
+    {
+      "epoch": 3.299168975069252,
+      "grad_norm": 8.01183795928955,
+      "learning_rate": 1.8263245819210745e-05,
+      "loss": 0.0632,
+      "step": 300
+    },
+    {
+      "epoch": 3.299168975069252,
+      "eval_accuracy": 0.9764052741151977,
+      "eval_f1": 0.9762893004716909,
+      "eval_loss": 0.08674201369285583,
+      "eval_precision": 0.9765146969702094,
+      "eval_recall": 0.9764052741151977,
+      "eval_runtime": 10.4868,
+      "eval_samples_per_second": 137.411,
+      "eval_steps_per_second": 17.26,
+      "step": 300
+    },
+    {
+      "epoch": 3.853185595567867,
+      "grad_norm": 18.632911682128906,
+      "learning_rate": 1.726071352087451e-05,
+      "loss": 0.0765,
+      "step": 350
+    },
+    {
+      "epoch": 4.3988919667590025,
+      "grad_norm": 3.3034017086029053,
+      "learning_rate": 1.609818420534627e-05,
+      "loss": 0.064,
+      "step": 400
+    },
+    {
+      "epoch": 4.3988919667590025,
+      "eval_accuracy": 0.9826509368494102,
+      "eval_f1": 0.9826055074466569,
+      "eval_loss": 0.0677182748913765,
+      "eval_precision": 0.9826109101098727,
+      "eval_recall": 0.9826509368494102,
+      "eval_runtime": 11.5916,
+      "eval_samples_per_second": 124.314,
+      "eval_steps_per_second": 15.615,
+      "step": 400
+    },
+    {
+      "epoch": 4.9529085872576175,
+      "grad_norm": 0.8745527267456055,
+      "learning_rate": 1.475799302451547e-05,
+      "loss": 0.0352,
+      "step": 450
+    },
+    {
+      "epoch": 5.498614958448753,
+      "grad_norm": 15.552165985107422,
+      "learning_rate": 1.3296319164004646e-05,
+      "loss": 0.0614,
+      "step": 500
+    },
+    {
+      "epoch": 5.498614958448753,
+      "eval_accuracy": 0.9826509368494102,
+      "eval_f1": 0.9827236461969681,
+      "eval_loss": 0.07660220563411713,
+      "eval_precision": 0.9829795881792385,
+      "eval_recall": 0.9826509368494102,
+      "eval_runtime": 11.5363,
+      "eval_samples_per_second": 124.91,
+      "eval_steps_per_second": 15.69,
+      "step": 500
+    },
+    {
+      "epoch": 6.044321329639889,
+      "grad_norm": 8.493371963500977,
+      "learning_rate": 1.1750482572892781e-05,
+      "loss": 0.0504,
+      "step": 550
+    },
+    {
+      "epoch": 6.598337950138504,
+      "grad_norm": 5.444666385650635,
+      "learning_rate": 1.0159952071429953e-05,
+      "loss": 0.0361,
+      "step": 600
+    },
+    {
+      "epoch": 6.598337950138504,
+      "eval_accuracy": 0.9764052741151977,
+      "eval_f1": 0.9762853272761889,
+      "eval_loss": 0.10010161995887756,
+      "eval_precision": 0.9763734760239566,
+      "eval_recall": 0.9764052741151977,
+      "eval_runtime": 11.3884,
+      "eval_samples_per_second": 126.532,
+      "eval_steps_per_second": 15.893,
+      "step": 600
+    },
+    {
+      "epoch": 7.14404432132964,
+      "grad_norm": 2.581049680709839,
+      "learning_rate": 8.565337619831516e-06,
+      "loss": 0.0297,
+      "step": 650
+    },
+    {
+      "epoch": 7.698060941828254,
+      "grad_norm": 0.20306643843650818,
+      "learning_rate": 7.007353451102557e-06,
+      "loss": 0.0321,
+      "step": 700
+    },
+    {
+      "epoch": 7.698060941828254,
+      "eval_accuracy": 0.9826509368494102,
+      "eval_f1": 0.9826866377257144,
+      "eval_loss": 0.07007652521133423,
+      "eval_precision": 0.982752646764885,
+      "eval_recall": 0.9826509368494102,
+      "eval_runtime": 11.7663,
+      "eval_samples_per_second": 122.468,
+      "eval_steps_per_second": 15.383,
+      "step": 700
+    },
+    {
+      "epoch": 8.24376731301939,
+      "grad_norm": 4.626258850097656,
+      "learning_rate": 5.52577854153435e-06,
+      "loss": 0.0344,
+      "step": 750
+    },
+    {
+      "epoch": 8.797783933518005,
+      "grad_norm": 0.3179256021976471,
+      "learning_rate": 4.184436663838245e-06,
+      "loss": 0.0495,
+      "step": 800
+    },
+    {
+      "epoch": 8.797783933518005,
+      "eval_accuracy": 0.9861207494795281,
+      "eval_f1": 0.9861160209367208,
+      "eval_loss": 0.06481500715017319,
+      "eval_precision": 0.9862368013315075,
+      "eval_recall": 0.9861207494795281,
+      "eval_runtime": 11.6949,
+      "eval_samples_per_second": 123.217,
+      "eval_steps_per_second": 15.477,
+      "step": 800
+    },
+    {
+      "epoch": 9.34349030470914,
+      "grad_norm": 11.409143447875977,
+      "learning_rate": 2.9629458799268897e-06,
+      "loss": 0.026,
+      "step": 850
+    },
+    {
+      "epoch": 9.897506925207756,
+      "grad_norm": 10.00415325164795,
+      "learning_rate": 1.921127530833138e-06,
+      "loss": 0.0339,
+      "step": 900
+    },
+    {
+      "epoch": 9.897506925207756,
+      "eval_accuracy": 0.9854267869535045,
+      "eval_f1": 0.9854149021422898,
+      "eval_loss": 0.06154085695743561,
+      "eval_precision": 0.9854658526912934,
+      "eval_recall": 0.9854267869535045,
+      "eval_runtime": 10.7933,
+      "eval_samples_per_second": 133.509,
+      "eval_steps_per_second": 16.77,
+      "step": 900
+    },
+    {
+      "epoch": 10.443213296398891,
+      "grad_norm": 23.903621673583984,
+      "learning_rate": 1.085581673374142e-06,
+      "loss": 0.0218,
+      "step": 950
+    },
+    {
+      "epoch": 10.997229916897506,
+      "grad_norm": 3.228843927383423,
+      "learning_rate": 4.776417456690152e-07,
+      "loss": 0.0384,
+      "step": 1000
+    },
+    {
+      "epoch": 10.997229916897506,
+      "eval_accuracy": 0.9854267869535045,
+      "eval_f1": 0.9854256575661703,
+      "eval_loss": 0.06275035440921783,
+      "eval_precision": 0.9854572946717711,
+      "eval_recall": 0.9854267869535045,
+      "eval_runtime": 11.6474,
+      "eval_samples_per_second": 123.718,
+      "eval_steps_per_second": 15.54,
+      "step": 1000
+    },
+    {
+      "epoch": 11.542936288088642,
+      "grad_norm": 3.3184566497802734,
+      "learning_rate": 1.1282987461486927e-07,
+      "loss": 0.024,
+      "step": 1050
+    },
+    {
+      "epoch": 12.0,
+      "step": 1092,
+      "total_flos": 3.1692355393112064e+18,
+      "train_loss": 0.048116779152726954,
+      "train_runtime": 1311.0723,
+      "train_samples_per_second": 52.729,
+      "train_steps_per_second": 0.833
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 1092,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 12,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.1692355393112064e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}