🔄 Incremental importance | Acc: 0.786, F1: 0.775

README.md

@@ -21,19 +21,19 @@ Fine-tuned FinBERT model for financial importance analysis in TradePulse.
 
 ## Performance
 
-*Last training: 2025-07-25 06:57*  
-*Dataset: `news_20250725.csv` (27 samples)*
+*Last training: 2025-07-25 07:33*  
+*Dataset: `news_20250725.csv` (125 samples)*
 
 | Metric | Value |
 |--------|-------|
-| Loss | 0.5277 |
-| Accuracy | 0.8571 |
-| F1 Score | 0.8707 |
+| Loss | 1.2877 |
+| Accuracy | 0.7188 |
+| F1 Score | 0.7188 |
 
-| F1 Macro | 0.8707 |
+| F1 Macro | 0.7188 |
 
-| Precision | 0.9286 |
-| Recall | 0.8571 |
+| Precision | 0.7188 |
+| Recall | 0.7188 |
 
 ## Training Details
 
@@ -65,4 +65,4 @@ predictions = outputs.logits.softmax(dim=-1)
 ## Model Card Authors
 
 - TradePulse ML Team
-- Auto-generated on 2025-07-25 06:57:56
+- Auto-generated on 2025-07-25 07:33:53

{checkpoint-10 → checkpoint-48}/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d01a6724d1bc83b694667d3ffff6f6bb77e7c7b5d6e326a0deaca2893ba205a
+oid sha256:4d69d2782f210b046e6bb1d1553c3b0120b093bba04194cc4f0245b5e8002062
 size 439039996

{checkpoint-10 → checkpoint-48}/trainer_state.json

@@ -1,53 +1,74 @@
 {
-  "best_metric": 0.870748299319728,
-  "best_model_checkpoint": "hf-importance-production/checkpoint-10",
+  "best_metric": 0.71875,
+  "best_model_checkpoint": "hf-importance-production/checkpoint-48",
   "epoch": 2.0,
   "eval_steps": 500,
-  "global_step": 10,
+  "global_step": 48,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
+    {
+      "epoch": 0.4166666666666667,
+      "grad_norm": 0.12729063630104065,
+      "learning_rate": 4.347826086956522e-06,
+      "loss": 0.0005,
+      "step": 10
+    },
+    {
+      "epoch": 0.8333333333333334,
+      "grad_norm": 0.023386986926198006,
+      "learning_rate": 8.695652173913044e-06,
+      "loss": 0.0004,
+      "step": 20
+    },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.7142857142857143,
-      "eval_f1": 0.7047619047619048,
-      "eval_f1_macro": 0.7111111111111111,
-      "eval_loss": 0.43314722180366516,
-      "eval_precision": 0.8333333333333333,
-      "eval_precision_macro": 0.7222222222222222,
-      "eval_recall": 0.7142857142857143,
-      "eval_recall_macro": 0.8333333333333334,
-      "eval_runtime": 3.9542,
-      "eval_samples_per_second": 1.77,
-      "eval_steps_per_second": 0.506,
-      "step": 5
+      "eval_accuracy": 0.6875,
+      "eval_f1": 0.6842334494773519,
+      "eval_f1_macro": 0.45063879210220675,
+      "eval_loss": 1.3590199947357178,
+      "eval_precision": 0.6824404761904761,
+      "eval_precision_macro": 0.45396825396825397,
+      "eval_recall": 0.6875,
+      "eval_recall_macro": 0.4484848484848485,
+      "eval_runtime": 17.8239,
+      "eval_samples_per_second": 1.795,
+      "eval_steps_per_second": 0.449,
+      "step": 24
     },
     {
-      "epoch": 2.0,
-      "grad_norm": 0.5402925610542297,
-      "learning_rate": 0.0,
-      "loss": 0.285,
-      "step": 10
+      "epoch": 1.25,
+      "grad_norm": 0.14419300854206085,
+      "learning_rate": 7.2000000000000005e-06,
+      "loss": 0.0002,
+      "step": 30
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.024391531944274902,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 0.0009,
+      "step": 40
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8571428571428571,
-      "eval_f1": 0.870748299319728,
-      "eval_f1_macro": 0.8412698412698413,
-      "eval_loss": 0.5276828408241272,
-      "eval_precision": 0.9285714285714286,
-      "eval_precision_macro": 0.8333333333333334,
-      "eval_recall": 0.8571428571428571,
-      "eval_recall_macro": 0.9166666666666666,
-      "eval_runtime": 3.9185,
-      "eval_samples_per_second": 1.786,
-      "eval_steps_per_second": 0.51,
-      "step": 10
+      "eval_accuracy": 0.71875,
+      "eval_f1": 0.71875,
+      "eval_f1_macro": 0.4787878787878788,
+      "eval_loss": 1.287726879119873,
+      "eval_precision": 0.71875,
+      "eval_precision_macro": 0.4787878787878788,
+      "eval_recall": 0.71875,
+      "eval_recall_macro": 0.4787878787878788,
+      "eval_runtime": 17.778,
+      "eval_samples_per_second": 1.8,
+      "eval_steps_per_second": 0.45,
+      "step": 48
     }
   ],
   "logging_steps": 10,
-  "max_steps": 10,
+  "max_steps": 48,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 2,
   "save_steps": 500,
@@ -72,7 +93,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 10524536709120.0,
+  "total_flos": 48939095697408.0,
   "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 sha256:1d01a6724d1bc83b694667d3ffff6f6bb77e7c7b5d6e326a0deaca2893ba205a
+oid sha256:4d69d2782f210b046e6bb1d1553c3b0120b093bba04194cc4f0245b5e8002062
 size 439039996

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fe388da48683bddd7c46856d680946afceeef503e46feec65f61458ffe1ef4f0
+oid sha256:93be2cf0885ccccc46f3fa93fc94df6d8f955436f3fad71bd10c8ec497c8fae9
 size 5521