End of training

README.md

@@ -1,15 +1,20 @@
 ---
 library_name: transformers
+language:
+- jv
 license: apache-2.0
 base_model: openai/whisper-tiny
 tags:
+- whisper
+- javanese
+- asr
 - generated_from_trainer
 datasets:
 - jv_id_asr_split
 metrics:
 - wer
 model-index:
-- name: whisper-tiny-javanese-openslr-v2
+- name: Whisper Tiny Java
   results:
   - task:
       name: Automatic Speech Recognition
@@ -29,7 +34,7 @@ model-index:
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 should probably proofread and complete it, then remove this comment. -->
 
-# whisper-tiny-javanese-openslr-v2
+# Whisper Tiny Java
 
 This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the jv_id_asr_split dataset.
 It achieves the following results on the evaluation set:

all_results.json

@@ -1,21 +1,21 @@
 {
-    "epoch": 4.9922212618841835,
-    "eval_loss": 0.5638014078140259,
-    "eval_runtime": 4431.4054,
+    "epoch": 4.32152117545376,
+    "eval_loss": 0.2792496979236603,
+    "eval_runtime": 2784.0877,
     "eval_samples": 18504,
-    "eval_samples_per_second": 4.176,
-    "eval_steps_per_second": 0.261,
-    "eval_wer": 0.4824482518630893,
+    "eval_samples_per_second": 6.646,
+    "eval_steps_per_second": 0.416,
+    "eval_wer": 0.6471586421539112,
     "pretrained_eval_loss": 4.0292205810546875,
     "pretrained_eval_model_preparation_time": 0.0019,
     "pretrained_eval_runtime": 3181.6206,
     "pretrained_eval_samples_per_second": 5.816,
     "pretrained_eval_steps_per_second": 0.364,
     "pretrained_eval_wer": 1.3043423861619188,
-    "total_flos": 1.819709817421824e+19,
-    "train_loss": 0.0,
-    "train_runtime": 0.166,
+    "total_flos": 1.574545111842816e+19,
+    "train_loss": 0.17131386260986328,
+    "train_runtime": 60097.0332,
     "train_samples": 148052,
-    "train_samples_per_second": 4460643.449,
-    "train_steps_per_second": 17414.502
+    "train_samples_per_second": 10.649,
+    "train_steps_per_second": 0.042
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 4.9922212618841835,
-    "eval_loss": 0.5638014078140259,
-    "eval_runtime": 4431.4054,
+    "epoch": 4.32152117545376,
+    "eval_loss": 0.2792496979236603,
+    "eval_runtime": 2784.0877,
     "eval_samples": 18504,
-    "eval_samples_per_second": 4.176,
-    "eval_steps_per_second": 0.261,
-    "eval_wer": 0.4824482518630893
+    "eval_samples_per_second": 6.646,
+    "eval_steps_per_second": 0.416,
+    "eval_wer": 0.6471586421539112
 }

runs/Mar16_11-27-13_dgx-a100/events.out.tfevents.1742162147.dgx-a100.690596.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e21f99fd4b547a7a4a11673fb97d703230d36858f63e11d5e9466d541fd28a3d
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 4.9922212618841835,
-    "total_flos": 1.819709817421824e+19,
-    "train_loss": 0.0,
-    "train_runtime": 0.166,
+    "epoch": 4.32152117545376,
+    "total_flos": 1.574545111842816e+19,
+    "train_loss": 0.17131386260986328,
+    "train_runtime": 60097.0332,
     "train_samples": 148052,
-    "train_samples_per_second": 4460643.449,
-    "train_steps_per_second": 17414.502
+    "train_samples_per_second": 10.649,
+    "train_steps_per_second": 0.042
 }

trainer_state.json

@@ -1,266 +1,246 @@
 {
-  "best_metric": 0.4824482518630893,
+  "best_global_step": null,
+  "best_metric": 0.47698996707825364,
   "best_model_checkpoint": "/raid/p-storage/slp01_565a7357/bagas-fine-tune-whisper/whisper-tiny-javanese-openslr-v2/checkpoint-500",
-  "epoch": 4.9922212618841835,
+  "epoch": 4.32152117545376,
   "eval_steps": 500,
-  "global_step": 2890,
+  "global_step": 2500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.17286084701815038,
-      "grad_norm": 6.286952018737793,
-      "learning_rate": 3.3910034602076125e-06,
-      "loss": 3.0341,
+      "grad_norm": 5.079843521118164,
+      "learning_rate": 7.840000000000001e-06,
+      "loss": 2.5618,
       "step": 100
     },
     {
       "epoch": 0.34572169403630076,
-      "grad_norm": 4.7146077156066895,
-      "learning_rate": 6.85121107266436e-06,
-      "loss": 1.3208,
+      "grad_norm": 4.103299617767334,
+      "learning_rate": 1.584e-05,
+      "loss": 1.021,
       "step": 200
     },
     {
       "epoch": 0.5185825410544511,
-      "grad_norm": 4.3300933837890625,
-      "learning_rate": 9.965397923875434e-06,
-      "loss": 0.9403,
+      "grad_norm": 3.8726439476013184,
+      "learning_rate": 1.9573333333333335e-05,
+      "loss": 0.7363,
       "step": 300
     },
     {
       "epoch": 0.6914433880726015,
-      "grad_norm": 4.19991397857666,
-      "learning_rate": 9.580930411380239e-06,
-      "loss": 0.7674,
+      "grad_norm": 3.9386720657348633,
+      "learning_rate": 1.8684444444444446e-05,
+      "loss": 0.6066,
       "step": 400
     },
     {
       "epoch": 0.8643042350907519,
-      "grad_norm": 3.640639066696167,
-      "learning_rate": 9.196462898885045e-06,
-      "loss": 0.6676,
+      "grad_norm": 3.273592472076416,
+      "learning_rate": 1.7795555555555557e-05,
+      "loss": 0.528,
       "step": 500
     },
     {
       "epoch": 0.8643042350907519,
-      "eval_loss": 0.5638014078140259,
-      "eval_runtime": 6063.2097,
-      "eval_samples_per_second": 3.052,
-      "eval_steps_per_second": 0.191,
-      "eval_wer": 0.4824482518630893,
+      "eval_loss": 0.4466875493526459,
+      "eval_runtime": 5073.2528,
+      "eval_samples_per_second": 3.647,
+      "eval_steps_per_second": 0.228,
+      "eval_wer": 0.47698996707825364,
       "step": 500
     },
     {
       "epoch": 1.0363007778738116,
-      "grad_norm": 3.6083388328552246,
-      "learning_rate": 8.811995386389852e-06,
-      "loss": 0.6009,
+      "grad_norm": 3.0979788303375244,
+      "learning_rate": 1.690666666666667e-05,
+      "loss": 0.4728,
       "step": 600
     },
     {
       "epoch": 1.2091616248919619,
-      "grad_norm": 3.529407024383545,
-      "learning_rate": 8.427527873894657e-06,
-      "loss": 0.5501,
+      "grad_norm": 3.1844944953918457,
+      "learning_rate": 1.601777777777778e-05,
+      "loss": 0.4222,
       "step": 700
     },
     {
       "epoch": 1.3820224719101124,
-      "grad_norm": 3.4671249389648438,
-      "learning_rate": 8.043060361399463e-06,
-      "loss": 0.523,
+      "grad_norm": 2.981661319732666,
+      "learning_rate": 1.5128888888888891e-05,
+      "loss": 0.404,
       "step": 800
     },
     {
       "epoch": 1.5548833189282627,
-      "grad_norm": 3.357516288757324,
-      "learning_rate": 7.658592848904268e-06,
-      "loss": 0.4992,
+      "grad_norm": 3.146735906600952,
+      "learning_rate": 1.4240000000000001e-05,
+      "loss": 0.3867,
       "step": 900
     },
     {
       "epoch": 1.727744165946413,
-      "grad_norm": 3.260972738265991,
-      "learning_rate": 7.274125336409074e-06,
-      "loss": 0.4792,
+      "grad_norm": 2.873340368270874,
+      "learning_rate": 1.3351111111111112e-05,
+      "loss": 0.3702,
       "step": 1000
     },
     {
       "epoch": 1.727744165946413,
-      "eval_loss": 0.42836251854896545,
-      "eval_runtime": 5157.0546,
-      "eval_samples_per_second": 3.588,
-      "eval_steps_per_second": 0.224,
-      "eval_wer": 0.5330351347758427,
+      "eval_loss": 0.34238535165786743,
+      "eval_runtime": 6401.1103,
+      "eval_samples_per_second": 2.891,
+      "eval_steps_per_second": 0.181,
+      "eval_wer": 0.5528429218245372,
       "step": 1000
     },
     {
-      "epoch": 1.9006050129645635,
-      "grad_norm": 3.132084608078003,
-      "learning_rate": 6.88965782391388e-06,
-      "loss": 0.4654,
+      "epoch": 1.902333621434745,
+      "grad_norm": 2.7809109687805176,
+      "learning_rate": 1.2462222222222222e-05,
+      "loss": 0.3611,
       "step": 1100
     },
     {
-      "epoch": 2.072601555747623,
-      "grad_norm": 3.0205070972442627,
-      "learning_rate": 6.505190311418685e-06,
-      "loss": 0.4387,
+      "epoch": 2.076058772687986,
+      "grad_norm": 2.62384295463562,
+      "learning_rate": 1.1573333333333335e-05,
+      "loss": 0.3371,
       "step": 1200
     },
     {
-      "epoch": 2.2454624027657735,
-      "grad_norm": 3.090315103530884,
-      "learning_rate": 6.120722798923492e-06,
-      "loss": 0.4201,
+      "epoch": 2.2489196197061365,
+      "grad_norm": 2.793154716491699,
+      "learning_rate": 1.0684444444444446e-05,
+      "loss": 0.3134,
       "step": 1300
     },
     {
-      "epoch": 2.4183232497839238,
-      "grad_norm": 3.003020763397217,
-      "learning_rate": 5.736255286428297e-06,
-      "loss": 0.4111,
+      "epoch": 2.421780466724287,
+      "grad_norm": 2.5597171783447266,
+      "learning_rate": 9.795555555555556e-06,
+      "loss": 0.3069,
       "step": 1400
     },
     {
-      "epoch": 2.591184096802074,
-      "grad_norm": 2.9050354957580566,
-      "learning_rate": 5.351787773933102e-06,
-      "loss": 0.3988,
+      "epoch": 2.5946413137424376,
+      "grad_norm": 2.6140079498291016,
+      "learning_rate": 8.906666666666667e-06,
+      "loss": 0.2988,
       "step": 1500
     },
     {
-      "epoch": 2.591184096802074,
-      "eval_loss": 0.3772418200969696,
-      "eval_runtime": 3820.8123,
-      "eval_samples_per_second": 4.843,
-      "eval_steps_per_second": 0.303,
-      "eval_wer": 0.5687485826445312,
+      "epoch": 2.5946413137424376,
+      "eval_loss": 0.30312925577163696,
+      "eval_runtime": 2962.721,
+      "eval_samples_per_second": 6.246,
+      "eval_steps_per_second": 0.391,
+      "eval_wer": 0.5552201690660703,
       "step": 1500
     },
     {
-      "epoch": 2.764044943820225,
-      "grad_norm": 2.983358383178711,
-      "learning_rate": 4.967320261437909e-06,
-      "loss": 0.392,
+      "epoch": 2.767502160760588,
+      "grad_norm": 2.635817050933838,
+      "learning_rate": 8.017777777777779e-06,
+      "loss": 0.2948,
       "step": 1600
     },
     {
-      "epoch": 2.936905790838375,
-      "grad_norm": 2.932623863220215,
-      "learning_rate": 4.5828527489427145e-06,
-      "loss": 0.3882,
+      "epoch": 2.940363007778738,
+      "grad_norm": 2.918149709701538,
+      "learning_rate": 7.12888888888889e-06,
+      "loss": 0.2903,
       "step": 1700
     },
     {
-      "epoch": 3.1089023336214345,
-      "grad_norm": 2.8963847160339355,
-      "learning_rate": 4.19838523644752e-06,
-      "loss": 0.3684,
+      "epoch": 3.1123595505617976,
+      "grad_norm": 2.2907419204711914,
+      "learning_rate": 6.24e-06,
+      "loss": 0.2711,
       "step": 1800
     },
     {
-      "epoch": 3.2817631806395853,
-      "grad_norm": 3.005558967590332,
-      "learning_rate": 3.8139177239523264e-06,
-      "loss": 0.3638,
+      "epoch": 3.2852203975799483,
+      "grad_norm": 2.640625238418579,
+      "learning_rate": 5.351111111111112e-06,
+      "loss": 0.2667,
       "step": 1900
     },
     {
-      "epoch": 3.4546240276577356,
-      "grad_norm": 2.9087977409362793,
-      "learning_rate": 3.429450211457132e-06,
-      "loss": 0.3565,
+      "epoch": 3.4580812445980986,
+      "grad_norm": 2.5180225372314453,
+      "learning_rate": 4.462222222222223e-06,
+      "loss": 0.2607,
       "step": 2000
     },
     {
-      "epoch": 3.4546240276577356,
-      "eval_loss": 0.35282832384109497,
-      "eval_runtime": 3674.1258,
-      "eval_samples_per_second": 5.036,
-      "eval_steps_per_second": 0.315,
-      "eval_wer": 0.6203598714409715,
+      "epoch": 3.4580812445980986,
+      "eval_loss": 0.28587546944618225,
+      "eval_runtime": 2830.646,
+      "eval_samples_per_second": 6.537,
+      "eval_steps_per_second": 0.409,
+      "eval_wer": 0.6484880238350316,
       "step": 2000
     },
     {
-      "epoch": 3.627484874675886,
-      "grad_norm": 2.949126958847046,
-      "learning_rate": 3.044982698961938e-06,
-      "loss": 0.3582,
+      "epoch": 3.630942091616249,
+      "grad_norm": 2.6452524662017822,
+      "learning_rate": 3.5733333333333336e-06,
+      "loss": 0.2637,
       "step": 2100
     },
     {
-      "epoch": 3.800345721694036,
-      "grad_norm": 2.861595869064331,
-      "learning_rate": 2.660515186466744e-06,
-      "loss": 0.3549,
+      "epoch": 3.8038029386343992,
+      "grad_norm": 2.3364474773406982,
+      "learning_rate": 2.6844444444444445e-06,
+      "loss": 0.2615,
       "step": 2200
     },
     {
-      "epoch": 3.973206568712187,
-      "grad_norm": 2.6879312992095947,
-      "learning_rate": 2.2760476739715497e-06,
-      "loss": 0.351,
+      "epoch": 3.9766637856525495,
+      "grad_norm": 2.414165496826172,
+      "learning_rate": 1.7955555555555556e-06,
+      "loss": 0.2603,
       "step": 2300
     },
     {
-      "epoch": 4.145203111495246,
-      "grad_norm": 3.0249099731445312,
-      "learning_rate": 1.8915801614763554e-06,
-      "loss": 0.339,
+      "epoch": 4.148660328435609,
+      "grad_norm": 2.3572444915771484,
+      "learning_rate": 9.066666666666668e-07,
+      "loss": 0.2483,
       "step": 2400
     },
     {
-      "epoch": 4.318063958513397,
-      "grad_norm": 2.900782585144043,
-      "learning_rate": 1.5071126489811613e-06,
-      "loss": 0.3386,
+      "epoch": 4.32152117545376,
+      "grad_norm": 2.6238534450531006,
+      "learning_rate": 1.777777777777778e-08,
+      "loss": 0.2481,
       "step": 2500
     },
     {
-      "epoch": 4.318063958513397,
-      "eval_loss": 0.34060564637184143,
-      "eval_runtime": 4820.2463,
-      "eval_samples_per_second": 3.839,
-      "eval_steps_per_second": 0.24,
-      "eval_wer": 0.6603038810125197,
+      "epoch": 4.32152117545376,
+      "eval_loss": 0.2792496979236603,
+      "eval_runtime": 2806.8606,
+      "eval_samples_per_second": 6.592,
+      "eval_steps_per_second": 0.412,
+      "eval_wer": 0.6471586421539112,
       "step": 2500
     },
     {
-      "epoch": 4.490924805531547,
-      "grad_norm": 2.7883245944976807,
-      "learning_rate": 1.122645136485967e-06,
-      "loss": 0.3351,
-      "step": 2600
-    },
-    {
-      "epoch": 4.663785652549698,
-      "grad_norm": 2.759507894515991,
-      "learning_rate": 7.381776239907729e-07,
-      "loss": 0.3364,
-      "step": 2700
-    },
-    {
-      "epoch": 4.8366464995678475,
-      "grad_norm": 2.8885769844055176,
-      "learning_rate": 3.5371011149557864e-07,
-      "loss": 0.334,
-      "step": 2800
-    },
-    {
-      "epoch": 4.9922212618841835,
-      "step": 2890,
-      "total_flos": 1.819709817421824e+19,
-      "train_loss": 0.0,
-      "train_runtime": 0.166,
-      "train_samples_per_second": 4460643.449,
-      "train_steps_per_second": 17414.502
+      "epoch": 4.32152117545376,
+      "step": 2500,
+      "total_flos": 1.574545111842816e+19,
+      "train_loss": 0.17131386260986328,
+      "train_runtime": 60097.0332,
+      "train_samples_per_second": 10.649,
+      "train_steps_per_second": 0.042
     }
   ],
   "logging_steps": 100,
-  "max_steps": 2890,
+  "max_steps": 2500,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 5,
   "save_steps": 500,
@@ -276,7 +256,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.819709817421824e+19,
+  "total_flos": 1.574545111842816e+19,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null