End of training

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 2.0,
-    "eval_loss": 0.2534329295158386,
-    "eval_runtime": 132.1673,
-    "eval_samples": 1549,
-    "eval_samples_per_second": 11.72,
-    "eval_steps_per_second": 1.468,
-    "eval_wer": 0.8469980632666236,
-    "train_loss": 2.1141159319460474,
-    "train_runtime": 3966.7132,
-    "train_samples": 11892,
-    "train_samples_per_second": 5.996,
-    "train_steps_per_second": 0.375
+    "eval_loss": 0.6070845127105713,
+    "eval_runtime": 277.1948,
+    "eval_samples": 1684,
+    "eval_samples_per_second": 6.075,
+    "eval_steps_per_second": 0.761,
+    "eval_wer": 0.9916864608076009,
+    "train_loss": 3.814326035825512,
+    "train_runtime": 2899.9163,
+    "train_samples": 5053,
+    "train_samples_per_second": 3.485,
+    "train_steps_per_second": 0.218
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 2.0,
-    "eval_loss": 0.2534329295158386,
-    "eval_runtime": 132.1673,
-    "eval_samples": 1549,
-    "eval_samples_per_second": 11.72,
-    "eval_steps_per_second": 1.468,
-    "eval_wer": 0.8469980632666236
+    "eval_loss": 0.6070845127105713,
+    "eval_runtime": 277.1948,
+    "eval_samples": 1684,
+    "eval_samples_per_second": 6.075,
+    "eval_steps_per_second": 0.761,
+    "eval_wer": 0.9916864608076009
 }

runs/Nov22_08-17-11_14e70937d4dd/events.out.tfevents.1669112076.14e70937d4dd.926.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8e1dfc26fa3b815ae8a17d44372657af37d32dd4aed690329fc8b7833c6da2c
+size 358

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
-    "train_loss": 2.1141159319460474,
-    "train_runtime": 3966.7132,
-    "train_samples": 11892,
-    "train_samples_per_second": 5.996,
-    "train_steps_per_second": 0.375
+    "train_loss": 3.814326035825512,
+    "train_runtime": 2899.9163,
+    "train_samples": 5053,
+    "train_samples_per_second": 3.485,
+    "train_steps_per_second": 0.218
 }

trainer_state.json

@@ -1,127 +1,70 @@
 {
-  "best_metric": 0.2534329295158386,
-  "best_model_checkpoint": "./xls-r-300m-bemba-fullset/checkpoint-1000",
-  "epoch": 1.9993275050437123,
-  "global_step": 1486,
+  "best_metric": 0.6070845127105713,
+  "best_model_checkpoint": "./xls-r-300m-bemba-fullset/checkpoint-500",
+  "epoch": 2.0,
+  "global_step": 632,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.13,
-      "learning_rate": 1.47e-05,
-      "loss": 11.7638,
+      "epoch": 0.32,
+      "learning_rate": 1.9599999999999995e-05,
+      "loss": 11.9405,
       "step": 100
     },
     {
-      "epoch": 0.27,
-      "learning_rate": 2.97e-05,
-      "loss": 3.7497,
+      "epoch": 0.63,
+      "learning_rate": 3.96e-05,
+      "loss": 3.6915,
       "step": 200
     },
     {
-      "epoch": 0.4,
-      "learning_rate": 4.4699999999999996e-05,
-      "loss": 3.0287,
+      "epoch": 0.95,
+      "learning_rate": 5.959999999999999e-05,
+      "loss": 2.984,
       "step": 300
     },
     {
-      "epoch": 0.54,
-      "learning_rate": 5.97e-05,
-      "loss": 2.8222,
+      "epoch": 1.27,
+      "learning_rate": 7.959999999999998e-05,
+      "loss": 2.7834,
       "step": 400
     },
     {
-      "epoch": 0.67,
-      "learning_rate": 7.47e-05,
-      "loss": 2.268,
+      "epoch": 1.58,
+      "learning_rate": 9.96e-05,
+      "loss": 1.6564,
       "step": 500
     },
     {
-      "epoch": 0.67,
-      "eval_loss": 1.0355395078659058,
-      "eval_runtime": 140.0953,
-      "eval_samples_per_second": 11.057,
-      "eval_steps_per_second": 1.385,
-      "eval_wer": 0.9928986442866365,
+      "epoch": 1.58,
+      "eval_loss": 0.6070845127105713,
+      "eval_runtime": 282.4761,
+      "eval_samples_per_second": 5.962,
+      "eval_steps_per_second": 0.747,
+      "eval_wer": 0.9916864608076009,
       "step": 500
     },
     {
-      "epoch": 0.81,
-      "learning_rate": 8.969999999999998e-05,
-      "loss": 1.1597,
+      "epoch": 1.9,
+      "learning_rate": 0.0001196,
+      "loss": 0.8314,
       "step": 600
     },
-    {
-      "epoch": 0.94,
-      "learning_rate": 0.00010469999999999998,
-      "loss": 0.9172,
-      "step": 700
-    },
-    {
-      "epoch": 1.08,
-      "learning_rate": 0.0001197,
-      "loss": 0.8284,
-      "step": 800
-    },
-    {
-      "epoch": 1.21,
-      "learning_rate": 0.0001347,
-      "loss": 0.7667,
-      "step": 900
-    },
-    {
-      "epoch": 1.35,
-      "learning_rate": 0.00014969999999999998,
-      "loss": 0.7365,
-      "step": 1000
-    },
-    {
-      "epoch": 1.35,
-      "eval_loss": 0.2534329295158386,
-      "eval_runtime": 132.8836,
-      "eval_samples_per_second": 11.657,
-      "eval_steps_per_second": 1.46,
-      "eval_wer": 0.8469980632666236,
-      "step": 1000
-    },
-    {
-      "epoch": 1.48,
-      "learning_rate": 0.0001647,
-      "loss": 0.7227,
-      "step": 1100
-    },
-    {
-      "epoch": 1.61,
-      "learning_rate": 0.00017969999999999998,
-      "loss": 0.7068,
-      "step": 1200
-    },
-    {
-      "epoch": 1.75,
-      "learning_rate": 0.0001947,
-      "loss": 0.6875,
-      "step": 1300
-    },
-    {
-      "epoch": 1.88,
-      "learning_rate": 0.00020969999999999997,
-      "loss": 0.6798,
-      "step": 1400
-    },
     {
       "epoch": 2.0,
-      "step": 1486,
-      "total_flos": 4.4337322315078743e+18,
-      "train_loss": 2.1141159319460474,
-      "train_runtime": 3966.7132,
-      "train_samples_per_second": 5.996,
-      "train_steps_per_second": 0.375
+      "step": 632,
+      "total_flos": 2.9733883621978947e+18,
+      "train_loss": 3.814326035825512,
+      "train_runtime": 2899.9163,
+      "train_samples_per_second": 3.485,
+      "train_steps_per_second": 0.218
     }
   ],
-  "max_steps": 1486,
+  "max_steps": 632,
   "num_train_epochs": 2,
-  "total_flos": 4.4337322315078743e+18,
+  "total_flos": 2.9733883621978947e+18,
   "trial_name": null,
   "trial_params": null
 }