🔄 Incremental importance | Acc: 0.803, F1: 0.646

README.md

@@ -21,19 +21,19 @@ Fine-tuned FinBERT model for financial importance analysis in TradePulse.
 
 ## Performance
 
-*Last training: 2025-07-29 15:39*  
+*Last training: 2025-07-30 11:16*  
 *Dataset: `base_reference.csv` (637 samples)*
 
 | Metric | Value |
 |--------|-------|
-| Loss | 0.8320 |
-| Accuracy | 0.8063 |
-| F1 Score | 0.8051 |
+| Loss | 0.9162 |
+| Accuracy | 0.8000 |
+| F1 Score | 0.7975 |
 
-| F1 Macro | 0.8051 |
+| F1 Macro | 0.7975 |
 
-| Precision | 0.8047 |
-| Recall | 0.8063 |
+| Precision | 0.7976 |
+| Recall | 0.8000 |
 
 ## Training Details
 
@@ -65,4 +65,4 @@ predictions = outputs.logits.softmax(dim=-1)
 ## Model Card Authors
 
 - TradePulse ML Team
-- Auto-generated on 2025-07-29 15:39:17
+- Auto-generated on 2025-07-30 11:16:11

{checkpoint-240 → checkpoint-120}/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fdc01d91706d553799d11f3cb8130f8e1b17d70daf22d597f48fcefb01b77f14
+oid sha256:7a306203b60f295a91fb646b11a9144870ea0e3b21f0b0210637c5bfe66b95f5
 size 439039996

{checkpoint-240 → checkpoint-120}/trainer_state.json

@@ -1,210 +1,111 @@
 {
-  "best_metric": 0.8051317614424409,
-  "best_model_checkpoint": "hf-importance-production/checkpoint-240",
-  "epoch": 2.0,
+  "best_metric": 0.7974805825242718,
+  "best_model_checkpoint": "hf-importance-production/checkpoint-120",
+  "epoch": 1.0,
   "eval_steps": 500,
-  "global_step": 240,
+  "global_step": 120,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.08333333333333333,
-      "grad_norm": 1.5605429410934448,
+      "grad_norm": 0.09946297109127045,
       "learning_rate": 8.403361344537816e-07,
-      "loss": 0.2922,
+      "loss": 0.0896,
       "step": 10
     },
     {
       "epoch": 0.16666666666666666,
-      "grad_norm": 2.0631015300750732,
+      "grad_norm": 0.07774800062179565,
       "learning_rate": 1.6806722689075632e-06,
-      "loss": 0.2751,
+      "loss": 0.1151,
       "step": 20
     },
     {
       "epoch": 0.25,
-      "grad_norm": 26.44710350036621,
+      "grad_norm": 1.0400522947311401,
       "learning_rate": 2.521008403361345e-06,
-      "loss": 0.1431,
+      "loss": 0.0195,
       "step": 30
     },
     {
       "epoch": 0.3333333333333333,
-      "grad_norm": 15.7521390914917,
+      "grad_norm": 0.2196376472711563,
       "learning_rate": 3.3613445378151265e-06,
-      "loss": 0.1394,
+      "loss": 0.001,
       "step": 40
     },
     {
       "epoch": 0.4166666666666667,
-      "grad_norm": 0.29773807525634766,
+      "grad_norm": 0.04618404060602188,
       "learning_rate": 4.201680672268908e-06,
-      "loss": 0.0197,
+      "loss": 0.0011,
       "step": 50
     },
     {
       "epoch": 0.5,
-      "grad_norm": 22.171756744384766,
+      "grad_norm": 0.020002124831080437,
       "learning_rate": 5.04201680672269e-06,
-      "loss": 0.0483,
+      "loss": 0.0008,
       "step": 60
     },
     {
       "epoch": 0.5833333333333334,
-      "grad_norm": 19.108678817749023,
+      "grad_norm": 0.7232803702354431,
       "learning_rate": 5.882352941176471e-06,
-      "loss": 0.1641,
+      "loss": 0.0056,
       "step": 70
     },
     {
       "epoch": 0.6666666666666666,
-      "grad_norm": 0.05635490268468857,
+      "grad_norm": 0.014796342700719833,
       "learning_rate": 6.722689075630253e-06,
-      "loss": 0.0468,
+      "loss": 0.0002,
       "step": 80
     },
     {
       "epoch": 0.75,
-      "grad_norm": 0.11963564902544022,
+      "grad_norm": 0.059538453817367554,
       "learning_rate": 7.563025210084034e-06,
-      "loss": 0.009,
+      "loss": 0.0008,
       "step": 90
     },
     {
       "epoch": 0.8333333333333334,
-      "grad_norm": 42.91386413574219,
+      "grad_norm": 6.586078643798828,
       "learning_rate": 8.403361344537815e-06,
-      "loss": 0.0908,
+      "loss": 0.0036,
       "step": 100
     },
     {
       "epoch": 0.9166666666666666,
-      "grad_norm": 1.6929007768630981,
+      "grad_norm": 0.012406314723193645,
       "learning_rate": 9.243697478991598e-06,
-      "loss": 0.1529,
+      "loss": 0.011,
       "step": 110
     },
     {
       "epoch": 1.0,
-      "grad_norm": 9.662502270657569e-05,
+      "grad_norm": 0.0004408100212458521,
       "learning_rate": 9.917355371900828e-06,
-      "loss": 0.0054,
+      "loss": 0.0002,
       "step": 120
     },
     {
       "epoch": 1.0,
       "eval_accuracy": 0.8,
-      "eval_f1": 0.7951267793019712,
-      "eval_f1_macro": 0.7037467600127446,
-      "eval_loss": 0.8872003555297852,
-      "eval_precision": 0.7960396600228876,
-      "eval_precision_macro": 0.7721565261618667,
+      "eval_f1": 0.7974805825242718,
+      "eval_f1_macro": 0.7152658344891355,
+      "eval_loss": 0.9162373542785645,
+      "eval_precision": 0.797591328783621,
+      "eval_precision_macro": 0.7225248991607884,
       "eval_recall": 0.8,
-      "eval_recall_macro": 0.6659389038634321,
-      "eval_runtime": 90.131,
-      "eval_samples_per_second": 1.775,
-      "eval_steps_per_second": 0.444,
+      "eval_recall_macro": 0.7102246181491464,
+      "eval_runtime": 88.5376,
+      "eval_samples_per_second": 1.807,
+      "eval_steps_per_second": 0.452,
       "step": 120
-    },
-    {
-      "epoch": 1.0833333333333333,
-      "grad_norm": 7.114067077636719,
-      "learning_rate": 9.090909090909091e-06,
-      "loss": 0.0037,
-      "step": 130
-    },
-    {
-      "epoch": 1.1666666666666667,
-      "grad_norm": 3.9886481761932373,
-      "learning_rate": 8.264462809917356e-06,
-      "loss": 0.014,
-      "step": 140
-    },
-    {
-      "epoch": 1.25,
-      "grad_norm": 2.7626068592071533,
-      "learning_rate": 7.43801652892562e-06,
-      "loss": 0.1729,
-      "step": 150
-    },
-    {
-      "epoch": 1.3333333333333333,
-      "grad_norm": 0.5616143345832825,
-      "learning_rate": 6.611570247933885e-06,
-      "loss": 0.0071,
-      "step": 160
-    },
-    {
-      "epoch": 1.4166666666666667,
-      "grad_norm": 1.5667345523834229,
-      "learning_rate": 5.785123966942149e-06,
-      "loss": 0.0452,
-      "step": 170
-    },
-    {
-      "epoch": 1.5,
-      "grad_norm": 9.732388496398926,
-      "learning_rate": 4.958677685950414e-06,
-      "loss": 0.0161,
-      "step": 180
-    },
-    {
-      "epoch": 1.5833333333333335,
-      "grad_norm": 1.119588851928711,
-      "learning_rate": 4.132231404958678e-06,
-      "loss": 0.0546,
-      "step": 190
-    },
-    {
-      "epoch": 1.6666666666666665,
-      "grad_norm": 0.037918105721473694,
-      "learning_rate": 3.3057851239669424e-06,
-      "loss": 0.1803,
-      "step": 200
-    },
-    {
-      "epoch": 1.75,
-      "grad_norm": 2.7715115547180176,
-      "learning_rate": 2.479338842975207e-06,
-      "loss": 0.4811,
-      "step": 210
-    },
-    {
-      "epoch": 1.8333333333333335,
-      "grad_norm": 68.59700775146484,
-      "learning_rate": 1.6528925619834712e-06,
-      "loss": 0.191,
-      "step": 220
-    },
-    {
-      "epoch": 1.9166666666666665,
-      "grad_norm": 9.923303604125977,
-      "learning_rate": 8.264462809917356e-07,
-      "loss": 0.157,
-      "step": 230
-    },
-    {
-      "epoch": 2.0,
-      "grad_norm": 4.033025470562279e-05,
-      "learning_rate": 0.0,
-      "loss": 0.2956,
-      "step": 240
-    },
-    {
-      "epoch": 2.0,
-      "eval_accuracy": 0.80625,
-      "eval_f1": 0.8051317614424409,
-      "eval_f1_macro": 0.7221451687471104,
-      "eval_loss": 0.8320282101631165,
-      "eval_precision": 0.8046626213592234,
-      "eval_precision_macro": 0.7253629218677761,
-      "eval_recall": 0.80625,
-      "eval_recall_macro": 0.7194699011680145,
-      "eval_runtime": 90.1147,
-      "eval_samples_per_second": 1.776,
-      "eval_steps_per_second": 0.444,
-      "step": 240
     }
   ],
   "logging_steps": 10,
@@ -228,12 +129,12 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": true
+        "should_training_stop": false
       },
       "attributes": {}
     }
   },
-  "total_flos": 251010200512512.0,
+  "total_flos": 125505100256256.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fdc01d91706d553799d11f3cb8130f8e1b17d70daf22d597f48fcefb01b77f14
+oid sha256:7a306203b60f295a91fb646b11a9144870ea0e3b21f0b0210637c5bfe66b95f5
 size 439039996