End of training

README.md

@@ -3,9 +3,23 @@ license: apache-2.0
 base_model: google/mt5-small
 tags:
 - generated_from_trainer
+datasets:
+- cwchang/tw_address_large
+metrics:
+- rouge
 model-index:
 - name: formatted_addr
-  results: []
+  results:
+  - task:
+      name: Summarization
+      type: summarization
+    dataset:
+      name: cwchang/tw_address_large
+      type: cwchang/tw_address_large
+    metrics:
+    - name: Rouge1
+      type: rouge
+      value: 96.9865
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -13,7 +27,14 @@ should probably proofread and complete it, then remove this comment. -->
 
 # formatted_addr
 
-This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on an unknown dataset.
+This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the cwchang/tw_address_large dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1493
+- Rouge1: 96.9865
+- Rouge2: 48.0164
+- Rougel: 96.9783
+- Rougelsum: 96.9638
+- Gen Len: 13.6825
 
 ## Model description
 

all_results.json

@@ -1,18 +1,18 @@
 {
     "epoch": 3.0,
-    "eval_gen_len": 13.375,
-    "eval_loss": 0.17634712159633636,
-    "eval_rouge1": 99.0667,
-    "eval_rouge2": 39.6667,
-    "eval_rougeL": 99.1333,
-    "eval_rougeLsum": 99.1333,
-    "eval_runtime": 16.2949,
-    "eval_samples": 200,
-    "eval_samples_per_second": 12.274,
-    "eval_steps_per_second": 3.068,
-    "train_loss": 2.2646766764322916,
-    "train_runtime": 282.894,
-    "train_samples": 2000,
-    "train_samples_per_second": 21.209,
-    "train_steps_per_second": 5.302
+    "eval_gen_len": 13.6825,
+    "eval_loss": 0.14928926527500153,
+    "eval_rouge1": 96.9865,
+    "eval_rouge2": 48.0164,
+    "eval_rougeL": 96.9783,
+    "eval_rougeLsum": 96.9638,
+    "eval_runtime": 148.8488,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 13.436,
+    "eval_steps_per_second": 3.359,
+    "train_loss": 0.4487843953450521,
+    "train_runtime": 2923.2046,
+    "train_samples": 20000,
+    "train_samples_per_second": 20.525,
+    "train_steps_per_second": 5.131
 }

eval_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 3.0,
-    "eval_gen_len": 13.375,
-    "eval_loss": 0.17634712159633636,
-    "eval_rouge1": 99.0667,
-    "eval_rouge2": 39.6667,
-    "eval_rougeL": 99.1333,
-    "eval_rougeLsum": 99.1333,
-    "eval_runtime": 16.2949,
-    "eval_samples": 200,
-    "eval_samples_per_second": 12.274,
-    "eval_steps_per_second": 3.068
+    "eval_gen_len": 13.6825,
+    "eval_loss": 0.14928926527500153,
+    "eval_rouge1": 96.9865,
+    "eval_rouge2": 48.0164,
+    "eval_rougeL": 96.9783,
+    "eval_rougeLsum": 96.9638,
+    "eval_runtime": 148.8488,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 13.436,
+    "eval_steps_per_second": 3.359
 }

generation_config.json

@@ -1,5 +1,4 @@
 {
-  "_from_model_config": true,
   "decoder_start_token_id": 0,
   "eos_token_id": 1,
   "pad_token_id": 0,

runs/Dec18_08-26-50_379bb1de15ca/events.out.tfevents.1702891185.379bb1de15ca.1410.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68a5795e90c00cfb5fb6f483bf329e2cc9623db1052abba3c452dd171868199a
+size 613

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.0,
-    "train_loss": 2.2646766764322916,
-    "train_runtime": 282.894,
-    "train_samples": 2000,
-    "train_samples_per_second": 21.209,
-    "train_steps_per_second": 5.302
+    "train_loss": 0.4487843953450521,
+    "train_runtime": 2923.2046,
+    "train_samples": 20000,
+    "train_samples_per_second": 20.525,
+    "train_steps_per_second": 5.131
 }

trainer_state.json

@@ -3,45 +3,207 @@
   "best_model_checkpoint": null,
   "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 1500,
+  "global_step": 15000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.8333333333333334e-05,
+      "loss": 5.4972,
+      "step": 500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.666666666666667e-05,
+      "loss": 0.9864,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.5e-05,
+      "loss": 0.6486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.3333333333333334e-05,
+      "loss": 0.4434,
+      "step": 2000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 0.4089,
+      "step": 2500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4e-05,
+      "loss": 0.3371,
+      "step": 3000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.8333333333333334e-05,
+      "loss": 0.3095,
+      "step": 3500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.6666666666666666e-05,
+      "loss": 0.2993,
+      "step": 4000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.5e-05,
+      "loss": 0.2893,
+      "step": 4500
+    },
     {
       "epoch": 1.0,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 5.4127,
-      "step": 500
+      "loss": 0.2869,
+      "step": 5000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.1666666666666666e-05,
+      "loss": 0.2181,
+      "step": 5500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3e-05,
+      "loss": 0.2417,
+      "step": 6000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8333333333333335e-05,
+      "loss": 0.2489,
+      "step": 6500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 0.226,
+      "step": 7000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.5e-05,
+      "loss": 0.2155,
+      "step": 7500
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 0.2154,
+      "step": 8000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.1666666666666667e-05,
+      "loss": 0.2066,
+      "step": 8500
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 2e-05,
+      "loss": 0.2167,
+      "step": 9000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.8333333333333333e-05,
+      "loss": 0.1876,
+      "step": 9500
     },
     {
       "epoch": 2.0,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.8232,
-      "step": 1000
+      "loss": 0.192,
+      "step": 10000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.5e-05,
+      "loss": 0.183,
+      "step": 10500
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 0.1728,
+      "step": 11000
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.1666666666666668e-05,
+      "loss": 0.18,
+      "step": 11500
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1e-05,
+      "loss": 0.1911,
+      "step": 12000
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.1742,
+      "step": 12500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 0.1834,
+      "step": 13000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 5e-06,
+      "loss": 0.1853,
+      "step": 13500
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 0.1756,
+      "step": 14000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 0.1615,
+      "step": 14500
     },
     {
       "epoch": 3.0,
       "learning_rate": 0.0,
-      "loss": 0.5581,
-      "step": 1500
+      "loss": 0.1816,
+      "step": 15000
     },
     {
       "epoch": 3.0,
-      "step": 1500,
-      "total_flos": 143865355345920.0,
-      "train_loss": 2.2646766764322916,
-      "train_runtime": 282.894,
-      "train_samples_per_second": 21.209,
-      "train_steps_per_second": 5.302
+      "step": 15000,
+      "total_flos": 1408713102213120.0,
+      "train_loss": 0.4487843953450521,
+      "train_runtime": 2923.2046,
+      "train_samples_per_second": 20.525,
+      "train_steps_per_second": 5.131
     }
   ],
   "logging_steps": 500,
-  "max_steps": 1500,
+  "max_steps": 15000,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 5000,
-  "total_flos": 143865355345920.0,
+  "total_flos": 1408713102213120.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null