🔄 Incremental importance | Acc: 0.775, F1: 0.644

README.md

@@ -21,19 +21,19 @@ Fine-tuned FinBERT model for financial importance analysis in TradePulse.
 
 ## Performance
 
-*Last training: 2025-07-29 13:33*  
+*Last training: 2025-07-29 15:39*  
 *Dataset: `base_reference.csv` (637 samples)*
 
 | Metric | Value |
 |--------|-------|
-| Loss | 0.6551 |
-| Accuracy | 0.8125 |
-| F1 Score | 0.8127 |
+| Loss | 0.8320 |
+| Accuracy | 0.8063 |
+| F1 Score | 0.8051 |
 
-| F1 Macro | 0.8127 |
+| F1 Macro | 0.8051 |
 
-| Precision | 0.8140 |
-| Recall | 0.8125 |
+| Precision | 0.8047 |
+| Recall | 0.8063 |
 
 ## Training Details
 
@@ -65,4 +65,4 @@ predictions = outputs.logits.softmax(dim=-1)
 ## Model Card Authors
 
 - TradePulse ML Team
-- Auto-generated on 2025-07-29 13:33:51
+- Auto-generated on 2025-07-29 15:39:17

checkpoint-240/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5adf2aaa4fe85784e8bc042af10862ff3a302c4a368f116c0a5f33220511fc72
+oid sha256:fdc01d91706d553799d11f3cb8130f8e1b17d70daf22d597f48fcefb01b77f14
 size 439039996

checkpoint-240/trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.8126807447116778,
+  "best_metric": 0.8051317614424409,
   "best_model_checkpoint": "hf-importance-production/checkpoint-240",
   "epoch": 2.0,
   "eval_steps": 500,
@@ -10,200 +10,200 @@
   "log_history": [
     {
       "epoch": 0.08333333333333333,
-      "grad_norm": 70.0009765625,
+      "grad_norm": 1.5605429410934448,
       "learning_rate": 8.403361344537816e-07,
-      "loss": 0.7483,
+      "loss": 0.2922,
       "step": 10
     },
     {
       "epoch": 0.16666666666666666,
-      "grad_norm": 24.547622680664062,
+      "grad_norm": 2.0631015300750732,
       "learning_rate": 1.6806722689075632e-06,
-      "loss": 0.9913,
+      "loss": 0.2751,
       "step": 20
     },
     {
       "epoch": 0.25,
-      "grad_norm": 57.52896499633789,
+      "grad_norm": 26.44710350036621,
       "learning_rate": 2.521008403361345e-06,
-      "loss": 0.6966,
+      "loss": 0.1431,
       "step": 30
     },
     {
       "epoch": 0.3333333333333333,
-      "grad_norm": 74.12036895751953,
+      "grad_norm": 15.7521390914917,
       "learning_rate": 3.3613445378151265e-06,
-      "loss": 0.6548,
+      "loss": 0.1394,
       "step": 40
     },
     {
       "epoch": 0.4166666666666667,
-      "grad_norm": 9.066349983215332,
+      "grad_norm": 0.29773807525634766,
       "learning_rate": 4.201680672268908e-06,
-      "loss": 0.3729,
+      "loss": 0.0197,
       "step": 50
     },
     {
       "epoch": 0.5,
-      "grad_norm": 50.364627838134766,
+      "grad_norm": 22.171756744384766,
       "learning_rate": 5.04201680672269e-06,
-      "loss": 0.6898,
+      "loss": 0.0483,
       "step": 60
     },
     {
       "epoch": 0.5833333333333334,
-      "grad_norm": 78.47008514404297,
+      "grad_norm": 19.108678817749023,
       "learning_rate": 5.882352941176471e-06,
-      "loss": 0.7254,
+      "loss": 0.1641,
       "step": 70
     },
     {
       "epoch": 0.6666666666666666,
-      "grad_norm": 0.9752311706542969,
+      "grad_norm": 0.05635490268468857,
       "learning_rate": 6.722689075630253e-06,
-      "loss": 0.6388,
+      "loss": 0.0468,
       "step": 80
     },
     {
       "epoch": 0.75,
-      "grad_norm": 12.438714027404785,
+      "grad_norm": 0.11963564902544022,
       "learning_rate": 7.563025210084034e-06,
-      "loss": 0.5719,
+      "loss": 0.009,
       "step": 90
     },
     {
       "epoch": 0.8333333333333334,
-      "grad_norm": 70.23839569091797,
+      "grad_norm": 42.91386413574219,
       "learning_rate": 8.403361344537815e-06,
-      "loss": 0.7578,
+      "loss": 0.0908,
       "step": 100
     },
     {
       "epoch": 0.9166666666666666,
-      "grad_norm": 85.30646514892578,
+      "grad_norm": 1.6929007768630981,
       "learning_rate": 9.243697478991598e-06,
-      "loss": 1.2087,
+      "loss": 0.1529,
       "step": 110
     },
     {
       "epoch": 1.0,
-      "grad_norm": 1.3721972703933716,
+      "grad_norm": 9.662502270657569e-05,
       "learning_rate": 9.917355371900828e-06,
-      "loss": 0.7927,
+      "loss": 0.0054,
       "step": 120
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.8125,
-      "eval_f1": 0.8083101323053494,
-      "eval_f1_macro": 0.6997588679484738,
-      "eval_loss": 0.6532053351402283,
-      "eval_precision": 0.8070379283788987,
-      "eval_precision_macro": 0.7347195481966371,
-      "eval_recall": 0.8125,
-      "eval_recall_macro": 0.6755615453728661,
-      "eval_runtime": 88.1395,
-      "eval_samples_per_second": 1.815,
-      "eval_steps_per_second": 0.454,
+      "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_recall": 0.8,
+      "eval_recall_macro": 0.6659389038634321,
+      "eval_runtime": 90.131,
+      "eval_samples_per_second": 1.775,
+      "eval_steps_per_second": 0.444,
       "step": 120
     },
     {
       "epoch": 1.0833333333333333,
-      "grad_norm": 53.19272232055664,
+      "grad_norm": 7.114067077636719,
       "learning_rate": 9.090909090909091e-06,
-      "loss": 0.347,
+      "loss": 0.0037,
       "step": 130
     },
     {
       "epoch": 1.1666666666666667,
-      "grad_norm": 130.35791015625,
+      "grad_norm": 3.9886481761932373,
       "learning_rate": 8.264462809917356e-06,
-      "loss": 0.6841,
+      "loss": 0.014,
       "step": 140
     },
     {
       "epoch": 1.25,
-      "grad_norm": 24.545459747314453,
+      "grad_norm": 2.7626068592071533,
       "learning_rate": 7.43801652892562e-06,
-      "loss": 0.5361,
+      "loss": 0.1729,
       "step": 150
     },
     {
       "epoch": 1.3333333333333333,
-      "grad_norm": 11.489201545715332,
+      "grad_norm": 0.5616143345832825,
       "learning_rate": 6.611570247933885e-06,
-      "loss": 0.1774,
+      "loss": 0.0071,
       "step": 160
     },
     {
       "epoch": 1.4166666666666667,
-      "grad_norm": 7.724330902099609,
+      "grad_norm": 1.5667345523834229,
       "learning_rate": 5.785123966942149e-06,
-      "loss": 0.1986,
+      "loss": 0.0452,
       "step": 170
     },
     {
       "epoch": 1.5,
-      "grad_norm": 51.44404602050781,
+      "grad_norm": 9.732388496398926,
       "learning_rate": 4.958677685950414e-06,
-      "loss": 0.2807,
+      "loss": 0.0161,
       "step": 180
     },
     {
       "epoch": 1.5833333333333335,
-      "grad_norm": 58.367225646972656,
+      "grad_norm": 1.119588851928711,
       "learning_rate": 4.132231404958678e-06,
-      "loss": 0.5279,
+      "loss": 0.0546,
       "step": 190
     },
     {
       "epoch": 1.6666666666666665,
-      "grad_norm": 0.692018985748291,
+      "grad_norm": 0.037918105721473694,
       "learning_rate": 3.3057851239669424e-06,
-      "loss": 0.3302,
+      "loss": 0.1803,
       "step": 200
     },
     {
       "epoch": 1.75,
-      "grad_norm": 39.534156799316406,
+      "grad_norm": 2.7715115547180176,
       "learning_rate": 2.479338842975207e-06,
-      "loss": 1.0246,
+      "loss": 0.4811,
       "step": 210
     },
     {
       "epoch": 1.8333333333333335,
-      "grad_norm": 45.754249572753906,
+      "grad_norm": 68.59700775146484,
       "learning_rate": 1.6528925619834712e-06,
-      "loss": 0.6398,
+      "loss": 0.191,
       "step": 220
     },
     {
       "epoch": 1.9166666666666665,
-      "grad_norm": 55.728824615478516,
+      "grad_norm": 9.923303604125977,
       "learning_rate": 8.264462809917356e-07,
-      "loss": 0.3741,
+      "loss": 0.157,
       "step": 230
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.00028416712302714586,
+      "grad_norm": 4.033025470562279e-05,
       "learning_rate": 0.0,
-      "loss": 0.4744,
+      "loss": 0.2956,
       "step": 240
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8125,
-      "eval_f1": 0.8126807447116778,
-      "eval_f1_macro": 0.6772336520815223,
-      "eval_loss": 0.6551325917243958,
-      "eval_precision": 0.8140347917079452,
-      "eval_precision_macro": 0.6750957664619245,
-      "eval_recall": 0.8125,
-      "eval_recall_macro": 0.6814734950584006,
-      "eval_runtime": 88.2081,
-      "eval_samples_per_second": 1.814,
-      "eval_steps_per_second": 0.453,
+      "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
     }
   ],

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5adf2aaa4fe85784e8bc042af10862ff3a302c4a368f116c0a5f33220511fc72
+oid sha256:fdc01d91706d553799d11f3cb8130f8e1b17d70daf22d597f48fcefb01b77f14
 size 439039996