End of training

README.md

@@ -1,12 +1,27 @@
 ---
+language:
+- en
 base_model: Hartunka/tiny_bert_km_100_v1
 tags:
 - generated_from_trainer
+datasets:
+- glue
 metrics:
 - accuracy
 model-index:
 - name: tiny_bert_km_100_v1_mnli
-  results: []
+  results:
+  - task:
+      name: Text Classification
+      type: text-classification
+    dataset:
+      name: GLUE MNLI
+      type: glue
+      args: mnli
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.6392392188771359
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -14,10 +29,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # tiny_bert_km_100_v1_mnli
 
-This model is a fine-tuned version of [Hartunka/tiny_bert_km_100_v1](https://huggingface.co/Hartunka/tiny_bert_km_100_v1) on an unknown dataset.
+This model is a fine-tuned version of [Hartunka/tiny_bert_km_100_v1](https://huggingface.co/Hartunka/tiny_bert_km_100_v1) on the GLUE MNLI dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.9526
-- Accuracy: 0.6473
+- Loss: 0.8203
+- Accuracy: 0.6392
 
 ## Model description
 

all_results.json

@@ -0,0 +1,22 @@
+{
+    "epoch": 10.0,
+    "epoch_mm": 10.0,
+    "eval_accuracy": 0.6323993886907794,
+    "eval_accuracy_mm": 0.6392392188771359,
+    "eval_loss": 0.8314237594604492,
+    "eval_loss_mm": 0.8203067183494568,
+    "eval_runtime": 3.4446,
+    "eval_runtime_mm": 3.5198,
+    "eval_samples": 9815,
+    "eval_samples_mm": 9832,
+    "eval_samples_per_second": 2849.412,
+    "eval_samples_per_second_mm": 2793.378,
+    "eval_steps_per_second": 11.322,
+    "eval_steps_per_second_mm": 11.08,
+    "total_flos": 1.029836100834816e+17,
+    "train_loss": 0.7496155541132762,
+    "train_runtime": 2650.3053,
+    "train_samples": 392702,
+    "train_samples_per_second": 7408.618,
+    "train_steps_per_second": 28.94
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 10.0,
+    "epoch_mm": 10.0,
+    "eval_accuracy": 0.6323993886907794,
+    "eval_accuracy_mm": 0.6392392188771359,
+    "eval_loss": 0.8314237594604492,
+    "eval_loss_mm": 0.8203067183494568,
+    "eval_runtime": 3.4446,
+    "eval_runtime_mm": 3.5198,
+    "eval_samples": 9815,
+    "eval_samples_mm": 9832,
+    "eval_samples_per_second": 2849.412,
+    "eval_samples_per_second_mm": 2793.378,
+    "eval_steps_per_second": 11.322,
+    "eval_steps_per_second_mm": 11.08
+}

logs/events.out.tfevents.1744824104.s_004_m

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b02a05b4d9bfac7d118f5b92ac2b2f1bbf42136cb42fbbbd397eee7a92172b5
+size 363

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 1.029836100834816e+17,
+    "train_loss": 0.7496155541132762,
+    "train_runtime": 2650.3053,
+    "train_samples": 392702,
+    "train_samples_per_second": 7408.618,
+    "train_steps_per_second": 28.94
+}

trainer_state.json

@@ -0,0 +1,190 @@
+{
+  "best_metric": 0.8314237594604492,
+  "best_model_checkpoint": "tiny_bert_km_100_v1_mnli/checkpoint-7670",
+  "epoch": 10.0,
+  "eval_steps": 500,
+  "global_step": 15340,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.0091803073883057,
+      "learning_rate": 4.9e-05,
+      "loss": 1.0013,
+      "step": 1534
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.5437595517065715,
+      "eval_loss": 0.9410763382911682,
+      "eval_runtime": 3.652,
+      "eval_samples_per_second": 2687.553,
+      "eval_steps_per_second": 10.679,
+      "step": 1534
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.4369374513626099,
+      "learning_rate": 4.8e-05,
+      "loss": 0.9118,
+      "step": 3068
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.5828833418237391,
+      "eval_loss": 0.8921328783035278,
+      "eval_runtime": 3.5758,
+      "eval_samples_per_second": 2744.85,
+      "eval_steps_per_second": 10.907,
+      "step": 3068
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.5986831188201904,
+      "learning_rate": 4.7e-05,
+      "loss": 0.8579,
+      "step": 4602
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.604890473764646,
+      "eval_loss": 0.8575275540351868,
+      "eval_runtime": 3.6158,
+      "eval_samples_per_second": 2714.455,
+      "eval_steps_per_second": 10.786,
+      "step": 4602
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 2.148224353790283,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 0.811,
+      "step": 6136
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.6191543555781966,
+      "eval_loss": 0.8428041934967041,
+      "eval_runtime": 3.6078,
+      "eval_samples_per_second": 2720.467,
+      "eval_steps_per_second": 10.81,
+      "step": 6136
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 1.6634807586669922,
+      "learning_rate": 4.5e-05,
+      "loss": 0.7625,
+      "step": 7670
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.6323993886907794,
+      "eval_loss": 0.8314237594604492,
+      "eval_runtime": 3.4027,
+      "eval_samples_per_second": 2884.439,
+      "eval_steps_per_second": 11.461,
+      "step": 7670
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 2.022226572036743,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 0.7165,
+      "step": 9204
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.6408558329088131,
+      "eval_loss": 0.8437779545783997,
+      "eval_runtime": 3.4697,
+      "eval_samples_per_second": 2828.773,
+      "eval_steps_per_second": 11.24,
+      "step": 9204
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 1.9079397916793823,
+      "learning_rate": 4.3e-05,
+      "loss": 0.6721,
+      "step": 10738
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.6440142638818136,
+      "eval_loss": 0.8497779369354248,
+      "eval_runtime": 3.4052,
+      "eval_samples_per_second": 2882.342,
+      "eval_steps_per_second": 11.453,
+      "step": 10738
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 2.122452735900879,
+      "learning_rate": 4.2e-05,
+      "loss": 0.6294,
+      "step": 12272
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.6461538461538462,
+      "eval_loss": 0.8689283132553101,
+      "eval_runtime": 3.4287,
+      "eval_samples_per_second": 2862.621,
+      "eval_steps_per_second": 11.375,
+      "step": 12272
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 2.5950520038604736,
+      "learning_rate": 4.1e-05,
+      "loss": 0.5874,
+      "step": 13806
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.6465613856342333,
+      "eval_loss": 0.9019529223442078,
+      "eval_runtime": 3.4591,
+      "eval_samples_per_second": 2837.477,
+      "eval_steps_per_second": 11.275,
+      "step": 13806
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 2.372058153152466,
+      "learning_rate": 4e-05,
+      "loss": 0.5462,
+      "step": 15340
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.6472745797249109,
+      "eval_loss": 0.9525629281997681,
+      "eval_runtime": 3.4343,
+      "eval_samples_per_second": 2857.97,
+      "eval_steps_per_second": 11.356,
+      "step": 15340
+    },
+    {
+      "epoch": 10.0,
+      "step": 15340,
+      "total_flos": 1.029836100834816e+17,
+      "train_loss": 0.7496155541132762,
+      "train_runtime": 2650.3053,
+      "train_samples_per_second": 7408.618,
+      "train_steps_per_second": 28.94
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 76700,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 50,
+  "save_steps": 500,
+  "total_flos": 1.029836100834816e+17,
+  "train_batch_size": 256,
+  "trial_name": null,
+  "trial_params": null
+}