End of training

README.md

@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [bioformers/bioformer-16L](https://huggingface.co/bioformers/bioformer-16L) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.2537
+- Loss: 2.0032
 
 ## Model description
 

all_results.json

@@ -1,9 +1,15 @@
 {
     "epoch": 10.0,
+    "eval_loss": 2.0032193660736084,
+    "eval_runtime": 1.0447,
+    "eval_samples": 951,
+    "eval_samples_per_second": 910.349,
+    "eval_steps_per_second": 57.435,
+    "perplexity": 7.412882507839082,
     "total_flos": 353880782323200.0,
-    "train_loss": 1.5547152099609376,
-    "train_runtime": 128.9033,
+    "train_loss": 1.5572135314941407,
+    "train_runtime": 127.1281,
     "train_samples": 2386,
-    "train_samples_per_second": 185.1,
-    "train_steps_per_second": 11.637
+    "train_samples_per_second": 187.685,
+    "train_steps_per_second": 11.799
 }

eval_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 10.0,
+    "eval_loss": 2.0032193660736084,
+    "eval_runtime": 1.0447,
+    "eval_samples": 951,
+    "eval_samples_per_second": 910.349,
+    "eval_steps_per_second": 57.435,
+    "perplexity": 7.412882507839082,
+    "total_flos": 353880782323200.0,
+    "train_loss": 1.5572135314941407,
+    "train_runtime": 127.1281,
+    "train_samples": 2386,
+    "train_samples_per_second": 187.685,
+    "train_steps_per_second": 11.799
+}

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 10.0,
     "total_flos": 353880782323200.0,
-    "train_loss": 1.5547152099609376,
-    "train_runtime": 128.9033,
+    "train_loss": 1.5572135314941407,
+    "train_runtime": 127.1281,
     "train_samples": 2386,
-    "train_samples_per_second": 185.1,
-    "train_steps_per_second": 11.637
+    "train_samples_per_second": 187.685,
+    "train_steps_per_second": 11.799
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 1.9935710430145264,
+  "best_metric": 1.9890172481536865,
   "best_model_checkpoint": "./TAPT-V1-Bioformer-16L/checkpoint-150",
   "epoch": 10.0,
   "eval_steps": 500,
@@ -10,162 +10,162 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 24.22607421875,
+      "grad_norm": 24.6746826171875,
       "learning_rate": 4.787234042553192e-05,
-      "loss": 2.0911,
+      "loss": 2.0908,
       "step": 150
     },
     {
       "epoch": 1.0,
-      "eval_loss": 1.9935710430145264,
-      "eval_runtime": 1.0627,
-      "eval_samples_per_second": 894.891,
-      "eval_steps_per_second": 56.46,
+      "eval_loss": 1.9890172481536865,
+      "eval_runtime": 1.0635,
+      "eval_samples_per_second": 894.184,
+      "eval_steps_per_second": 56.415,
       "step": 150
     },
     {
       "epoch": 2.0,
-      "grad_norm": 23.085124969482422,
+      "grad_norm": 22.7192440032959,
       "learning_rate": 4.2553191489361704e-05,
-      "loss": 1.8617,
+      "loss": 1.862,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_loss": 2.1117889881134033,
-      "eval_runtime": 1.0282,
-      "eval_samples_per_second": 924.919,
-      "eval_steps_per_second": 58.355,
+      "eval_loss": 2.1128599643707275,
+      "eval_runtime": 1.0114,
+      "eval_samples_per_second": 940.235,
+      "eval_steps_per_second": 59.321,
       "step": 300
     },
     {
       "epoch": 3.0,
-      "grad_norm": 48.75080108642578,
+      "grad_norm": 48.27261734008789,
       "learning_rate": 3.723404255319149e-05,
-      "loss": 1.6883,
+      "loss": 1.6906,
       "step": 450
     },
     {
       "epoch": 3.0,
-      "eval_loss": 2.146700620651245,
-      "eval_runtime": 1.0388,
-      "eval_samples_per_second": 915.474,
-      "eval_steps_per_second": 57.759,
+      "eval_loss": 2.1430840492248535,
+      "eval_runtime": 1.0129,
+      "eval_samples_per_second": 938.928,
+      "eval_steps_per_second": 59.238,
       "step": 450
     },
     {
       "epoch": 4.0,
-      "grad_norm": 22.47693634033203,
+      "grad_norm": 23.66407585144043,
       "learning_rate": 3.191489361702128e-05,
-      "loss": 1.56,
+      "loss": 1.5674,
       "step": 600
     },
     {
       "epoch": 4.0,
-      "eval_loss": 2.108654022216797,
-      "eval_runtime": 1.0009,
-      "eval_samples_per_second": 950.138,
-      "eval_steps_per_second": 59.946,
+      "eval_loss": 2.110779285430908,
+      "eval_runtime": 1.0204,
+      "eval_samples_per_second": 931.969,
+      "eval_steps_per_second": 58.799,
       "step": 600
     },
     {
       "epoch": 5.0,
-      "grad_norm": 25.13186264038086,
+      "grad_norm": 25.759119033813477,
       "learning_rate": 2.6595744680851064e-05,
-      "loss": 1.5605,
+      "loss": 1.5622,
       "step": 750
     },
     {
       "epoch": 5.0,
-      "eval_loss": 2.1210551261901855,
-      "eval_runtime": 1.0438,
-      "eval_samples_per_second": 911.089,
-      "eval_steps_per_second": 57.482,
+      "eval_loss": 2.124483108520508,
+      "eval_runtime": 1.04,
+      "eval_samples_per_second": 914.426,
+      "eval_steps_per_second": 57.692,
       "step": 750
     },
     {
       "epoch": 6.0,
-      "grad_norm": 42.86140441894531,
+      "grad_norm": 42.69515609741211,
       "learning_rate": 2.1276595744680852e-05,
-      "loss": 1.4688,
+      "loss": 1.4715,
       "step": 900
     },
     {
       "epoch": 6.0,
-      "eval_loss": 2.108290433883667,
-      "eval_runtime": 1.0161,
-      "eval_samples_per_second": 935.924,
-      "eval_steps_per_second": 59.049,
+      "eval_loss": 2.1117031574249268,
+      "eval_runtime": 1.0222,
+      "eval_samples_per_second": 930.375,
+      "eval_steps_per_second": 58.699,
       "step": 900
     },
     {
       "epoch": 7.0,
-      "grad_norm": 58.89534378051758,
+      "grad_norm": 64.62760925292969,
       "learning_rate": 1.595744680851064e-05,
-      "loss": 1.4312,
+      "loss": 1.4321,
       "step": 1050
     },
     {
       "epoch": 7.0,
-      "eval_loss": 2.1463418006896973,
-      "eval_runtime": 1.022,
-      "eval_samples_per_second": 930.534,
-      "eval_steps_per_second": 58.709,
+      "eval_loss": 2.157745599746704,
+      "eval_runtime": 1.032,
+      "eval_samples_per_second": 921.502,
+      "eval_steps_per_second": 58.139,
       "step": 1050
     },
     {
       "epoch": 8.0,
-      "grad_norm": 34.060401916503906,
+      "grad_norm": 26.836322784423828,
       "learning_rate": 1.0638297872340426e-05,
-      "loss": 1.316,
+      "loss": 1.3189,
       "step": 1200
     },
     {
       "epoch": 8.0,
-      "eval_loss": 2.1740951538085938,
-      "eval_runtime": 1.0283,
-      "eval_samples_per_second": 924.84,
-      "eval_steps_per_second": 58.35,
+      "eval_loss": 2.1733434200286865,
+      "eval_runtime": 1.0399,
+      "eval_samples_per_second": 914.531,
+      "eval_steps_per_second": 57.699,
       "step": 1200
     },
     {
       "epoch": 9.0,
-      "grad_norm": 37.21686935424805,
+      "grad_norm": 44.0683479309082,
       "learning_rate": 5.319148936170213e-06,
-      "loss": 1.2566,
+      "loss": 1.2598,
       "step": 1350
     },
     {
       "epoch": 9.0,
-      "eval_loss": 2.1824934482574463,
-      "eval_runtime": 1.0427,
-      "eval_samples_per_second": 912.046,
-      "eval_steps_per_second": 57.542,
+      "eval_loss": 2.178889036178589,
+      "eval_runtime": 1.0302,
+      "eval_samples_per_second": 923.107,
+      "eval_steps_per_second": 58.24,
       "step": 1350
     },
     {
       "epoch": 10.0,
-      "grad_norm": 78.66870880126953,
+      "grad_norm": 73.13053894042969,
       "learning_rate": 0.0,
-      "loss": 1.313,
+      "loss": 1.3167,
       "step": 1500
     },
     {
       "epoch": 10.0,
-      "eval_loss": 2.2559754848480225,
-      "eval_runtime": 1.0323,
-      "eval_samples_per_second": 921.223,
-      "eval_steps_per_second": 58.121,
+      "eval_loss": 2.2537429332733154,
+      "eval_runtime": 1.072,
+      "eval_samples_per_second": 887.127,
+      "eval_steps_per_second": 55.97,
       "step": 1500
     },
     {
       "epoch": 10.0,
       "step": 1500,
       "total_flos": 353880782323200.0,
-      "train_loss": 1.5547152099609376,
-      "train_runtime": 128.9033,
-      "train_samples_per_second": 185.1,
-      "train_steps_per_second": 11.637
+      "train_loss": 1.5572135314941407,
+      "train_runtime": 127.1281,
+      "train_samples_per_second": 187.685,
+      "train_steps_per_second": 11.799
     }
   ],
   "logging_steps": 500,