End of training

README.md

@@ -1,4 +1,6 @@
 ---
+language:
+- en
 license: apache-2.0
 tags:
 - generated_from_trainer
@@ -13,7 +15,7 @@ model-index:
       name: Text Classification
       type: text-classification
     dataset:
-      name: glue
+      name: GLUE QNLI
       type: glue
       config: qnli
       split: validation
@@ -21,7 +23,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.5742266154127769
+      value: 0.5881383855024712
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -29,10 +31,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # distilbert_sa_GLUE_Experiment_logit_kd_qnli_384
 
-This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset.
+This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the GLUE QNLI dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4226
-- Accuracy: 0.5742
+- Loss: 0.3912
+- Accuracy: 0.5881
 
 ## Model description
 

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 8.0,
+    "eval_accuracy": 0.5881383855024712,
+    "eval_loss": 0.39115190505981445,
+    "eval_runtime": 5.8451,
+    "eval_samples": 5463,
+    "eval_samples_per_second": 934.636,
+    "eval_steps_per_second": 3.764,
+    "train_loss": 0.36694961175685975,
+    "train_runtime": 1362.7354,
+    "train_samples": 104743,
+    "train_samples_per_second": 3843.116,
+    "train_steps_per_second": 15.043
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.0,
+    "eval_accuracy": 0.5881383855024712,
+    "eval_loss": 0.39115190505981445,
+    "eval_runtime": 5.8451,
+    "eval_samples": 5463,
+    "eval_samples_per_second": 934.636,
+    "eval_steps_per_second": 3.764
+}

logs/events.out.tfevents.1675031410.serv-3334.1026928.8

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42026bb4b45bb370a8b2bd51f4a79d91f255eaa894b5fb9b06ebeedd8ae548c1
+size 363

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 8.0,
+    "train_loss": 0.36694961175685975,
+    "train_runtime": 1362.7354,
+    "train_samples": 104743,
+    "train_samples_per_second": 3843.116,
+    "train_steps_per_second": 15.043
+}

trainer_state.json

@@ -0,0 +1,145 @@
+{
+  "best_metric": 0.39115190505981445,
+  "best_model_checkpoint": "distilbert_sa_GLUE_Experiment_logit_kd_qnli_384/checkpoint-1230",
+  "epoch": 8.0,
+  "global_step": 3280,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.9e-05,
+      "loss": 0.4059,
+      "step": 410
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.5733113673805601,
+      "eval_loss": 0.3930094242095947,
+      "eval_runtime": 5.8524,
+      "eval_samples_per_second": 933.462,
+      "eval_steps_per_second": 3.759,
+      "step": 410
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.8e-05,
+      "loss": 0.3918,
+      "step": 820
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.5839282445542742,
+      "eval_loss": 0.3918865919113159,
+      "eval_runtime": 5.8358,
+      "eval_samples_per_second": 936.114,
+      "eval_steps_per_second": 3.77,
+      "step": 820
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 4.7e-05,
+      "loss": 0.3807,
+      "step": 1230
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.5881383855024712,
+      "eval_loss": 0.39115190505981445,
+      "eval_runtime": 5.8838,
+      "eval_samples_per_second": 928.486,
+      "eval_steps_per_second": 3.739,
+      "step": 1230
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 0.371,
+      "step": 1640
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.5842943437671609,
+      "eval_loss": 0.3949229121208191,
+      "eval_runtime": 5.8416,
+      "eval_samples_per_second": 935.19,
+      "eval_steps_per_second": 3.766,
+      "step": 1640
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 4.5e-05,
+      "loss": 0.3618,
+      "step": 2050
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.5815485996705108,
+      "eval_loss": 0.3984927833080292,
+      "eval_runtime": 5.8237,
+      "eval_samples_per_second": 938.059,
+      "eval_steps_per_second": 3.778,
+      "step": 2050
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 0.352,
+      "step": 2460
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.5800842028189639,
+      "eval_loss": 0.41364026069641113,
+      "eval_runtime": 5.8333,
+      "eval_samples_per_second": 936.522,
+      "eval_steps_per_second": 3.771,
+      "step": 2460
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.3e-05,
+      "loss": 0.3416,
+      "step": 2870
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.5773384587223137,
+      "eval_loss": 0.42220625281333923,
+      "eval_runtime": 5.9263,
+      "eval_samples_per_second": 921.819,
+      "eval_steps_per_second": 3.712,
+      "step": 2870
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.2e-05,
+      "loss": 0.331,
+      "step": 3280
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.5742266154127769,
+      "eval_loss": 0.42263802886009216,
+      "eval_runtime": 5.8241,
+      "eval_samples_per_second": 937.998,
+      "eval_steps_per_second": 3.777,
+      "step": 3280
+    },
+    {
+      "epoch": 8.0,
+      "step": 3280,
+      "total_flos": 2.3017230102953984e+16,
+      "train_loss": 0.36694961175685975,
+      "train_runtime": 1362.7354,
+      "train_samples_per_second": 3843.116,
+      "train_steps_per_second": 15.043
+    }
+  ],
+  "max_steps": 20500,
+  "num_train_epochs": 50,
+  "total_flos": 2.3017230102953984e+16,
+  "trial_name": null,
+  "trial_params": null
+}