Train FinImpact Direction1D V4

README.md

@@ -0,0 +1,419 @@
+---
+library_name: transformers
+tags:
+- finance
+- sentiment-analysis
+- text-classification
+- market-impact
+- gated-fusion
+- multitask-learning
+- event-study
+datasets:
+- NickyNicky/finance-financialmodelingprep-stock-news-sentiments-rss-feed
+- siddharthmb/stocks-ohlcv
+license: mit
+---
+
+# FinImpact Direction1D V4
+
+This repository contains a custom Hugging Face/PyTorch model for financial text impact modeling.
+It combines a financial language encoder with market context features through a gated fusion head.
+This v4 checkpoint is trained primarily as a 1-day market-impact signal model and adds return-aware auxiliary losses for directional-score alignment and confidence calibration.
+
+## Task
+
+Input:
+
+- Financial news title/body
+- Symbol-level market context available before the event date
+
+Outputs:
+
+- `direction_1d`: bearish / neutral / bullish future 1-day abnormal market reaction
+- auxiliary diagnostics: `sentiment`, `direction_5d`, and `volatility` heads are still exported, but their default loss weights are zero in this v4 run.
+
+## Architecture
+
+- Encoder: `microsoft/deberta-v3-large`
+- Pooling: attention-mask-aware mean pooling over `last_hidden_state`
+- Text branch: LayerNorm -> Linear -> GELU -> Dropout
+- Numeric branch: LayerNorm -> Linear -> GELU -> Dropout
+- Fusion: learned sigmoid gate between text and numeric branches
+- Heads: 1D direction classification head plus auxiliary diagnostic heads
+- Return-aware training: `P(bullish) - P(bearish)` is softly aligned to realized 1-day abnormal return sign/magnitude
+
+This is intentionally not an embedding-only classifier. The model fine-tunes a contextual encoder and learns a supervised 1-day market-impact boundary.
+
+## Training Data
+
+News source:
+
+- `NickyNicky/finance-financialmodelingprep-stock-news-sentiments-rss-feed`
+- Fields used: `symbol`, `publishedDate`, `title`, `text`, `sentiment`, `sentimentScore`
+
+Price source:
+
+- `siddharthmb/stocks-ohlcv`
+- Daily close data used to build lagged market features and future-return labels.
+
+Training rows after join: `13930`
+
+Time split:
+
+- train: `10865`
+- validation: `1532`
+- test: `1533`
+
+Symbols:
+
+`AAPL, ABNB, ADBE, ADI, ADP, AMAT, AMD, AMZN, ARKK, ASML, AVGO, BA, BABA, BAC, BIDU, BMY, CAT, CMCSA, COIN, COST, CRM, CSCO, CVX, DHR, DIS, ELF, GE, GM, GOOGL, HD, IBM, INTC, IOVA, JNJ, JPM, KO, LULU, MA, MCD, MDB, META, MRNA, MSFT, MU, NFLX, NKE, NVDA, ORCL, PEP, PFE, PLTR, PYPL, QCOM, RCL, SHOP, SNOW, SPOT, TGT, TSLA, TSM, TXN, UAL, UBER, UNH, UPS, WMT, XOM, ZM`
+
+## Numeric Features
+
+- `source_sentiment_score`
+- `sentiment_abs_score`
+- `site_benzinga`
+- `site_globenewswire`
+- `site_yahoo_finance`
+- `site_marketwatch`
+- `text_char_len`
+- `title_char_len`
+- `is_after_close`
+- `is_premarket`
+- `is_market_hours`
+- `published_hour_sin`
+- `published_hour_cos`
+- `ret_1d_lag`
+- `ret_5d_lag`
+- `ret_20d_lag`
+- `vol_5d`
+- `vol_20d`
+- `vol_60d`
+- `vol_ratio_5_20`
+- `drawdown_20d`
+- `volume_z_20d`
+- `sector_code`
+- `market_ret_1d_lag`
+- `market_ret_5d_lag`
+- `sector_ret_1d_lag`
+- `sector_ret_5d_lag`
+- `beta_market_60d`
+- `beta_sector_60d`
+
+All numeric features are fit/scaled on the training split only.
+
+## Label Construction
+
+- `sentiment` comes from the source dataset label.
+- `direction_1d` is generated from future 1-day beta-adjusted abnormal log return after the event anchor date.
+- `direction_5d` is generated from future 5-day beta-adjusted abnormal log return.
+- A return is neutral when it falls inside a volatility-adjusted threshold.
+- `volatility` is based on train-split tertiles of future absolute 5-day return.
+
+Event anchoring shifts after-close UTC news to the next calendar day and uses the next available trading date.
+Duplicate same-symbol same-day event titles are removed before training.
+
+## Test Metrics
+
+```json
+{
+  "sentiment": {
+    "accuracy": 0.009784735812133072,
+    "macro_precision": 0.0032615786040443573,
+    "macro_recall": 0.3333333333333333,
+    "macro_f1": 0.006459948320413436,
+    "confusion_matrix": [
+      [
+        0,
+        110,
+        0
+      ],
+      [
+        0,
+        15,
+        0
+      ],
+      [
+        0,
+        1408,
+        0
+      ]
+    ],
+    "predicted_class_counts": {
+      "bearish": 0,
+      "neutral": 1533,
+      "bullish": 0
+    },
+    "true_class_counts": {
+      "bearish": 110,
+      "neutral": 15,
+      "bullish": 1408
+    },
+    "expected_calibration_error": 0.8247767172447622,
+    "brier_score": 1.5716914553389048
+  },
+  "direction_1d": {
+    "accuracy": 0.3333333333333333,
+    "macro_precision": 0.2797514006239241,
+    "macro_recall": 0.30693224852151285,
+    "macro_f1": 0.28541099148120913,
+    "confusion_matrix": [
+      [
+        31,
+        217,
+        212
+      ],
+      [
+        54,
+        138,
+        244
+      ],
+      [
+        124,
+        171,
+        342
+      ]
+    ],
+    "predicted_class_counts": {
+      "bearish": 209,
+      "neutral": 526,
+      "bullish": 798
+    },
+    "true_class_counts": {
+      "bearish": 460,
+      "neutral": 436,
+      "bullish": 637
+    },
+    "expected_calibration_error": 0.03496117605077468,
+    "brier_score": 0.6723385830557309
+  },
+  "direction_5d": {
+    "accuracy": 0.45857795172863663,
+    "macro_precision": 0.28120269843193446,
+    "macro_recall": 0.33741824434028733,
+    "macro_f1": 0.30491929241175314,
+    "confusion_matrix": [
+      [
+        0,
+        123,
+        143
+      ],
+      [
+        0,
+        553,
+        311
+      ],
+      [
+        0,
+        253,
+        150
+      ]
+    ],
+    "predicted_class_counts": {
+      "bearish": 0,
+      "neutral": 929,
+      "bullish": 604
+    },
+    "true_class_counts": {
+      "bearish": 266,
+      "neutral": 864,
+      "bullish": 403
+    },
+    "expected_calibration_error": 0.024249568105288305,
+    "brier_score": 0.633983250156182
+  },
+  "volatility": {
+    "accuracy": 0.39921722113502933,
+    "macro_precision": 0.13307240704500978,
+    "macro_recall": 0.3333333333333333,
+    "macro_f1": 0.1902097902097902,
+    "confusion_matrix": [
+      [
+        0,
+        534,
+        0
+      ],
+      [
+        0,
+        612,
+        0
+      ],
+      [
+        0,
+        387,
+        0
+      ]
+    ],
+    "predicted_class_counts": {
+      "lower_vol": 0,
+      "normal_vol": 1533,
+      "higher_vol": 0
+    },
+    "true_class_counts": {
+      "lower_vol": 534,
+      "normal_vol": 612,
+      "higher_vol": 387
+    },
+    "expected_calibration_error": 0.1760233100613036,
+    "brier_score": 0.712886643106662
+  },
+  "event_backtest": {
+    "split": "test",
+    "transaction_cost_bps": 10.0,
+    "active_1d_fraction": 0.6568819308545336,
+    "hit_ratio_1d_active": 0.5789473684210527,
+    "avg_signal_return_1d_all": 0.0023507131510749455,
+    "avg_signal_return_1d_all_net": 0.0016938312202204116,
+    "avg_signal_return_1d_active": 0.0035785931088360393,
+    "avg_signal_return_1d_active_net": 0.002578593108836038,
+    "avg_abnormal_signal_return_1d_all": 0.0007162017627715418,
+    "avg_abnormal_signal_return_1d_all_net": 5.93198319170076e-05,
+    "avg_abnormal_signal_return_1d_active": 0.0010903051661656142,
+    "avg_abnormal_signal_return_1d_active_net": 9.030516616561321e-05,
+    "soft_signal_backtest_1d": {
+      "mean_abs_signal": 0.049029190093278885,
+      "active_fraction_abs_score_ge_0.10": 0.11545988258317025,
+      "hit_ratio_all_gross": 0.532941943900848,
+      "hit_ratio_active_gross": 0.4463276836158192,
+      "avg_signal_return_all": -4.1903094825102016e-05,
+      "avg_signal_return_all_net": -9.093229164136574e-05,
+      "avg_abnormal_signal_return_all": 6.6492548285168596e-06,
+      "avg_abnormal_signal_return_all_net": -4.237994289724156e-05
+    },
+    "active_5d_fraction": 0.39399869536855836,
+    "hit_ratio_5d_active": 0.3708609271523179,
+    "avg_signal_return_5d_all": -0.004762698527096634,
+    "avg_signal_return_5d_active": -0.012088107354369436,
+    "confidence_backtests_1d": {
+      "top_10pct_directional_confidence": {
+        "num_events": 153,
+        "mean_confidence": 0.40472039580345154,
+        "hit_ratio": 0.48366013071895425,
+        "avg_signal_return": -0.0017291086948829782,
+        "avg_signal_return_net": -0.002729108694882979,
+        "median_signal_return": -0.0011537353275343776,
+        "avg_abnormal_signal_return": 0.0023775517351216227,
+        "avg_abnormal_signal_return_net": 0.001377551735121622
+      },
+      "top_20pct_directional_confidence": {
+        "num_events": 307,
+        "mean_confidence": 0.39327916502952576,
+        "hit_ratio": 0.46579804560260585,
+        "avg_signal_return": -0.0018346647964980064,
+        "avg_signal_return_net": -0.0028346647964980075,
+        "median_signal_return": -0.001494303229264915,
+        "avg_abnormal_signal_return": 0.00013626769358741198,
+        "avg_abnormal_signal_return_net": -0.0008637323064125889
+      },
+      "top_30pct_directional_confidence": {
+        "num_events": 460,
+        "mean_confidence": 0.3858344554901123,
+        "hit_ratio": 0.5021739130434782,
+        "avg_signal_return": 0.000168503345380684,
+        "avg_signal_return_net": -0.0008314966546193169,
+        "median_signal_return": 0.00047751690726727247,
+        "avg_abnormal_signal_return": -0.0006967808645482443,
+        "avg_abnormal_signal_return_net": -0.0016967808645482452
+      },
+      "threshold_0.34": {
+        "num_events": 1108,
+        "coverage": 0.7227658186562296,
+        "hit_ratio": 0.5523465703971119,
+        "avg_signal_return": 0.002400788538323842,
+        "avg_signal_return_net": 0.001400788538323841,
+        "median_signal_return": 0.0029742957558482885,
+        "avg_abnormal_signal_return": 0.00043566977963306825,
+        "avg_abnormal_signal_return_net": -0.0005643302203669326
+      },
+      "threshold_0.36": {
+        "num_events": 636,
+        "coverage": 0.41487279843444225,
+        "hit_ratio": 0.5471698113207547,
+        "avg_signal_return": 0.001725182487273244,
+        "avg_signal_return_net": 0.000725182487273243,
+        "median_signal_return": 0.0027859483379870653,
+        "avg_abnormal_signal_return": -0.0001395135721791607,
+        "avg_abnormal_signal_return_net": -0.0011395135721791617
+      },
+      "threshold_0.38": {
+        "num_events": 247,
+        "coverage": 0.16112198303979125,
+        "hit_ratio": 0.44129554655870445,
+        "avg_signal_return": -0.0028902063559107334,
+        "avg_signal_return_net": -0.0038902063559107343,
+        "median_signal_return": -0.0022296553943306208,
+        "avg_abnormal_signal_return": 0.0008437850360328761,
+        "avg_abnormal_signal_return_net": -0.00015621496396712474
+      },
+      "threshold_0.40": {
+        "num_events": 85,
+        "coverage": 0.055446836268754074,
+        "hit_ratio": 0.43529411764705883,
+        "avg_signal_return": -0.0034284473146887168,
+        "avg_signal_return_net": -0.004428447314688717,
+        "median_signal_return": -0.0031941309571266174,
+        "avg_abnormal_signal_return": 0.0008126156158087884,
+        "avg_abnormal_signal_return_net": -0.00018738438419121242
+      },
+      "threshold_0.42": {
+        "num_events": 20,
+        "coverage": 0.01304631441617743,
+        "hit_ratio": 0.25,
+        "avg_signal_return": -0.008969856356270612,
+        "avg_signal_return_net": -0.009969856356270613,
+        "median_signal_return": -0.011959618888795376,
+        "avg_abnormal_signal_return": -0.0006033775032847188,
+        "avg_abnormal_signal_return_net": -0.0016033775032847197
+      },
+      "threshold_0.45": {
+        "num_events": 2,
+        "coverage": 0.001304631441617743,
+        "hit_ratio": 0.5,
+        "avg_signal_return": 0.0010717622935771942,
+        "avg_signal_return_net": 7.176229357719333e-05,
+        "median_signal_return": 0.0010717622935771942,
+        "avg_abnormal_signal_return": 0.01302456425037235,
+        "avg_abnormal_signal_return_net": 0.01202456425037235
+      },
+      "threshold_0.50": {
+        "num_events": 0,
+        "coverage": 0.0,
+        "hit_ratio": null,
+        "avg_signal_return": null,
+        "avg_signal_return_net": null,
+        "median_signal_return": null,
+        "avg_abnormal_signal_return": null,
+        "avg_abnormal_signal_return_net": null
+      }
+    },
+    "mean_text_gate_weight": 0.506599485874176,
+    "std_text_gate_weight": 0.0016385371563956141
+  }
+}
+```
+
+The `confidence_backtests_1d` section ranks events by directional confidence, where directional confidence is `max(P(bearish), P(bullish))`.
+This is the preferred way to inspect whether the model is useful as a signal filter.
+V4 also reports cost-adjusted backtest fields using `10.0` bps per active directional event.
+
+## Important Limitations
+
+This is a research model, not trading advice.
+
+Known limitations:
+
+- Public news sentiment labels can be noisy.
+- Daily OHLCV alignment is an approximation; intraday timestamp alignment would be better.
+- Market-direction labels are derived from future returns and are sensitive to threshold choice.
+- Results should be evaluated out-of-sample by date and by ticker before any practical use.
+
+## Files
+
+- `pytorch_model.bin`: custom gated-fusion model weights.
+- `training_config.json`: training and dataset configuration.
+- `feature_schema.json`: numeric feature scaler and schema.
+- `label_mapping.json`: task label names.
+- `metrics.json`: train/validation/test metrics.
+- `test_predictions.csv`: event-level test predictions and returns.
+- `top_10pct_directional_confidence_1d` in `test_predictions.csv`: marks the highest-confidence directional event subset.
+- `confusion_matrix_*.csv`: confusion matrices per split/task.

class_weights.json

@@ -0,0 +1,22 @@
+{
+  "sentiment": [
+    0.3636363744735718,
+    2.2727272510528564,
+    0.3636363744735718
+  ],
+  "direction_1d": [
+    0.9880715608596802,
+    0.9986989498138428,
+    1.013229489326477
+  ],
+  "direction_5d": [
+    1.2122101783752441,
+    0.5105218887329102,
+    1.2772679328918457
+  ],
+  "volatility": [
+    1.0096524953842163,
+    1.0141775608062744,
+    0.9761698842048645
+  ]
+}

confusion_matrix_test_direction_1d.csv

@@ -0,0 +1,4 @@
+,bearish,neutral,bullish
+bearish,31,217,212
+neutral,54,138,244
+bullish,124,171,342

confusion_matrix_test_direction_5d.csv

@@ -0,0 +1,4 @@
+,bearish,neutral,bullish
+bearish,0,123,143
+neutral,0,553,311
+bullish,0,253,150

confusion_matrix_test_sentiment.csv

@@ -0,0 +1,4 @@
+,bearish,neutral,bullish
+bearish,0,110,0
+neutral,0,15,0
+bullish,0,1408,0

confusion_matrix_test_volatility.csv

@@ -0,0 +1,4 @@
+,lower_vol,normal_vol,higher_vol
+lower_vol,0,534,0
+normal_vol,0,612,0
+higher_vol,0,387,0

confusion_matrix_validation_direction_1d.csv

@@ -0,0 +1,4 @@
+,bearish,neutral,bullish
+bearish,104,200,164
+neutral,116,228,186
+bullish,154,189,191

confusion_matrix_validation_direction_5d.csv

@@ -0,0 +1,4 @@
+,bearish,neutral,bullish
+bearish,0,131,208
+neutral,0,381,424
+bullish,0,166,222

confusion_matrix_validation_sentiment.csv

@@ -0,0 +1,4 @@
+,bearish,neutral,bullish
+bearish,0,72,0
+neutral,0,13,0
+bullish,0,1447,0

confusion_matrix_validation_volatility.csv

@@ -0,0 +1,4 @@
+,lower_vol,normal_vol,higher_vol
+lower_vol,0,535,0
+normal_vol,0,642,0
+higher_vol,0,355,0

feature_schema.json

@@ -0,0 +1,95 @@
+{
+  "numeric_columns": [
+    "source_sentiment_score",
+    "sentiment_abs_score",
+    "site_benzinga",
+    "site_globenewswire",
+    "site_yahoo_finance",
+    "site_marketwatch",
+    "text_char_len",
+    "title_char_len",
+    "is_after_close",
+    "is_premarket",
+    "is_market_hours",
+    "published_hour_sin",
+    "published_hour_cos",
+    "ret_1d_lag",
+    "ret_5d_lag",
+    "ret_20d_lag",
+    "vol_5d",
+    "vol_20d",
+    "vol_60d",
+    "vol_ratio_5_20",
+    "drawdown_20d",
+    "volume_z_20d",
+    "sector_code",
+    "market_ret_1d_lag",
+    "market_ret_5d_lag",
+    "sector_ret_1d_lag",
+    "sector_ret_5d_lag",
+    "beta_market_60d",
+    "beta_sector_60d"
+  ],
+  "mean": {
+    "source_sentiment_score": 0.746147335480902,
+    "sentiment_abs_score": 0.8758751153945923,
+    "site_benzinga": 0.3414634168148041,
+    "site_globenewswire": 0.027059365063905716,
+    "site_yahoo_finance": 0.19834330677986145,
+    "site_marketwatch": 0.0011965024750679731,
+    "text_char_len": 0.12647317349910736,
+    "title_char_len": 0.22998590767383575,
+    "is_after_close": 0.06129774451255798,
+    "is_premarket": 0.4531063139438629,
+    "is_market_hours": 0.4855959415435791,
+    "published_hour_sin": -0.2567417025566101,
+    "published_hour_cos": -0.35876917839050293,
+    "ret_1d_lag": 0.0008439657480443655,
+    "ret_5d_lag": 0.005167760459005587,
+    "ret_20d_lag": 0.02857849583185001,
+    "vol_5d": 0.02562241014758805,
+    "vol_20d": 0.027127421020100354,
+    "vol_60d": 0.030643427833552454,
+    "vol_ratio_5_20": 0.94897371917211,
+    "drawdown_20d": -0.06130785621108734,
+    "volume_z_20d": 0.9134664162837107,
+    "sector_code": 0.2863782788771284,
+    "market_ret_1d_lag": 0.0006297079118320088,
+    "market_ret_5d_lag": 0.0041547514374490275,
+    "sector_ret_1d_lag": 0.0008193551641604897,
+    "sector_ret_5d_lag": 0.0056375516939092185,
+    "beta_market_60d": 1.3920874921870132,
+    "beta_sector_60d": 1.1672672796791717
+  },
+  "std": {
+    "source_sentiment_score": 0.5201509712378972,
+    "sentiment_abs_score": 0.24518661201000214,
+    "site_benzinga": 0.4742223620414734,
+    "site_globenewswire": 0.1622639149427414,
+    "site_yahoo_finance": 0.39877045154571533,
+    "site_marketwatch": 0.034571390599012375,
+    "text_char_len": 0.02897312119603157,
+    "title_char_len": 0.08457622677087784,
+    "is_after_close": 0.23988670110702515,
+    "is_premarket": 0.49781903624534607,
+    "is_market_hours": 0.4998154640197754,
+    "published_hour_sin": 0.663528323173523,
+    "published_hour_cos": 0.6042951941490173,
+    "ret_1d_lag": 0.043107957015871196,
+    "ret_5d_lag": 0.10119434487495654,
+    "ret_20d_lag": 0.19046219546936188,
+    "vol_5d": 0.03433107463832683,
+    "vol_20d": 0.025843168460955513,
+    "vol_60d": 0.03346308394931953,
+    "vol_ratio_5_20": 0.368944284961192,
+    "drawdown_20d": 0.09463532321589767,
+    "volume_z_20d": 2.680127570980756,
+    "sector_code": 0.266243410025217,
+    "market_ret_1d_lag": 0.009415043142061333,
+    "market_ret_5d_lag": 0.020352265466718426,
+    "sector_ret_1d_lag": 0.016764466116674252,
+    "sector_ret_5d_lag": 0.04006510290929962,
+    "beta_market_60d": 0.6201234265286986,
+    "beta_sector_60d": 0.5781415112568552
+  }
+}

label_mapping.json

@@ -0,0 +1,22 @@
+{
+  "sentiment": [
+    "bearish",
+    "neutral",
+    "bullish"
+  ],
+  "direction_1d": [
+    "bearish",
+    "neutral",
+    "bullish"
+  ],
+  "direction_5d": [
+    "bearish",
+    "neutral",
+    "bullish"
+  ],
+  "volatility": [
+    "lower_vol",
+    "normal_vol",
+    "higher_vol"
+  ]
+}

metrics.json

@@ -0,0 +1,640 @@
+{
+  "train": {
+    "train_runtime": 638.1554,
+    "train_samples_per_second": 51.077,
+    "train_steps_per_second": 3.197,
+    "total_flos": 0.0,
+    "train_loss": 0.6800105861589021,
+    "epoch": 3.0
+  },
+  "validation": {
+    "sentiment": {
+      "accuracy": 0.008485639686684074,
+      "macro_precision": 0.0028285465622280245,
+      "macro_recall": 0.3333333333333333,
+      "macro_f1": 0.0056094929881337656,
+      "confusion_matrix": [
+        [
+          0,
+          72,
+          0
+        ],
+        [
+          0,
+          13,
+          0
+        ],
+        [
+          0,
+          1447,
+          0
+        ]
+      ],
+      "predicted_class_counts": {
+        "bearish": 0,
+        "neutral": 1532,
+        "bullish": 0
+      },
+      "true_class_counts": {
+        "bearish": 72,
+        "neutral": 13,
+        "bullish": 1447
+      },
+      "expected_calibration_error": 0.8264312032931156,
+      "brier_score": 1.573439388720477
+    },
+    "direction_1d": {
+      "accuracy": 0.3413838120104439,
+      "macro_precision": 0.3335515858937544,
+      "macro_recall": 0.33669626802974273,
+      "macro_f1": 0.33331285508030406,
+      "confusion_matrix": [
+        [
+          104,
+          200,
+          164
+        ],
+        [
+          116,
+          228,
+          186
+        ],
+        [
+          154,
+          189,
+          191
+        ]
+      ],
+      "predicted_class_counts": {
+        "bearish": 374,
+        "neutral": 617,
+        "bullish": 541
+      },
+      "true_class_counts": {
+        "bearish": 468,
+        "neutral": 530,
+        "bullish": 534
+      },
+      "expected_calibration_error": 0.028834249269868946,
+      "brier_score": 0.6675477741251082
+    },
+    "direction_5d": {
+      "accuracy": 0.3936031331592689,
+      "macro_precision": 0.27396668808245767,
+      "macro_recall": 0.3484856246398156,
+      "macro_f1": 0.29043708459698375,
+      "confusion_matrix": [
+        [
+          0,
+          131,
+          208
+        ],
+        [
+          0,
+          381,
+          424
+        ],
+        [
+          0,
+          166,
+          222
+        ]
+      ],
+      "predicted_class_counts": {
+        "bearish": 0,
+        "neutral": 678,
+        "bullish": 854
+      },
+      "true_class_counts": {
+        "bearish": 339,
+        "neutral": 805,
+        "bullish": 388
+      },
+      "expected_calibration_error": 0.060579302202783736,
+      "brier_score": 0.6606167988480304
+    },
+    "volatility": {
+      "accuracy": 0.41906005221932113,
+      "macro_precision": 0.13968668407310705,
+      "macro_recall": 0.3333333333333333,
+      "macro_f1": 0.1968721251149954,
+      "confusion_matrix": [
+        [
+          0,
+          535,
+          0
+        ],
+        [
+          0,
+          642,
+          0
+        ],
+        [
+          0,
+          355,
+          0
+        ]
+      ],
+      "predicted_class_counts": {
+        "lower_vol": 0,
+        "normal_vol": 1532,
+        "higher_vol": 0
+      },
+      "true_class_counts": {
+        "lower_vol": 535,
+        "normal_vol": 642,
+        "higher_vol": 355
+      },
+      "expected_calibration_error": 0.15814054444159914,
+      "brier_score": 0.6913779982386878
+    },
+    "event_backtest": {
+      "split": "validation",
+      "transaction_cost_bps": 10.0,
+      "active_1d_fraction": 0.5972584856396866,
+      "hit_ratio_1d_active": 0.5398907103825137,
+      "avg_signal_return_1d_all": -0.000668341834270596,
+      "avg_signal_return_1d_all_net": -0.0012656003199102831,
+      "avg_signal_return_1d_active": -0.0011190160547568885,
+      "avg_signal_return_1d_active_net": -0.0021190160547568896,
+      "avg_abnormal_signal_return_1d_all": -0.0009099995935003474,
+      "avg_abnormal_signal_return_1d_all_net": -0.0015072580791400346,
+      "avg_abnormal_signal_return_1d_active": -0.001523627734691292,
+      "avg_abnormal_signal_return_1d_active_net": -0.002523627734691293,
+      "soft_signal_backtest_1d": {
+        "mean_abs_signal": 0.04358483850955963,
+        "active_fraction_abs_score_ge_0.10": 0.0731070496083551,
+        "hit_ratio_all_gross": 0.5071801566579635,
+        "hit_ratio_active_gross": 0.5446428571428571,
+        "avg_signal_return_all": -2.6098849048139527e-05,
+        "avg_signal_return_all_net": -6.968369416426867e-05,
+        "avg_abnormal_signal_return_all": -3.516719152685255e-05,
+        "avg_abnormal_signal_return_all_net": -7.87520402809605e-05
+      },
+      "active_5d_fraction": 0.5574412532637075,
+      "hit_ratio_5d_active": 0.5644028103044496,
+      "avg_signal_return_5d_all": 0.00222230488512584,
+      "avg_signal_return_5d_active": 0.003986617194394364,
+      "confidence_backtests_1d": {
+        "top_10pct_directional_confidence": {
+          "num_events": 153,
+          "mean_confidence": 0.39838218688964844,
+          "hit_ratio": 0.5424836601307189,
+          "avg_signal_return": -0.0040096313650019805,
+          "avg_signal_return_net": -0.005009631365001981,
+          "median_signal_return": 0.001708349445834756,
+          "avg_abnormal_signal_return": -0.0016484424157800104,
+          "avg_abnormal_signal_return_net": -0.002648442415780011
+        },
+        "top_20pct_directional_confidence": {
+          "num_events": 306,
+          "mean_confidence": 0.38756096363067627,
+          "hit_ratio": 0.5588235294117647,
+          "avg_signal_return": -0.0005135064179175138,
+          "avg_signal_return_net": -0.0015135064179175146,
+          "median_signal_return": 0.003223672043532133,
+          "avg_abnormal_signal_return": -0.0006505049740809268,
+          "avg_abnormal_signal_return_net": -0.0016505049740809278
+        },
+        "top_30pct_directional_confidence": {
+          "num_events": 460,
+          "mean_confidence": 0.38047993183135986,
+          "hit_ratio": 0.5543478260869565,
+          "avg_signal_return": -0.0009027605532142131,
+          "avg_signal_return_net": -0.001902760553214214,
+          "median_signal_return": 0.002927212161011994,
+          "avg_abnormal_signal_return": -0.0014120966409751892,
+          "avg_abnormal_signal_return_net": -0.0024120966409751903
+        },
+        "threshold_0.34": {
+          "num_events": 1040,
+          "coverage": 0.6788511749347258,
+          "hit_ratio": 0.5394230769230769,
+          "avg_signal_return": -0.0007571742408449959,
+          "avg_signal_return_net": -0.0017571742408449967,
+          "median_signal_return": 0.0012542950571514666,
+          "avg_abnormal_signal_return": -0.0012730416878412143,
+          "avg_abnormal_signal_return_net": -0.0022730416878412154
+        },
+        "threshold_0.36": {
+          "num_events": 494,
+          "coverage": 0.3224543080939948,
+          "hit_ratio": 0.5566801619433198,
+          "avg_signal_return": -0.0008215918299352735,
+          "avg_signal_return_net": -0.0018215918299352744,
+          "median_signal_return": 0.0024513842072337866,
+          "avg_abnormal_signal_return": -0.0013882744387551298,
+          "avg_abnormal_signal_return_net": -0.002388274438755131
+        },
+        "threshold_0.38": {
+          "num_events": 188,
+          "coverage": 0.1227154046997389,
+          "hit_ratio": 0.5478723404255319,
+          "avg_signal_return": -0.0020706070304669917,
+          "avg_signal_return_net": -0.0030706070304669926,
+          "median_signal_return": 0.0030422902200371027,
+          "avg_abnormal_signal_return": -0.0007841120187801239,
+          "avg_abnormal_signal_return_net": -0.0017841120187801248
+        },
+        "threshold_0.40": {
+          "num_events": 49,
+          "coverage": 0.03198433420365535,
+          "hit_ratio": 0.5714285714285714,
+          "avg_signal_return": -0.0023946920445435966,
+          "avg_signal_return_net": -0.0033946920445435975,
+          "median_signal_return": 0.003524972125887871,
+          "avg_abnormal_signal_return": 0.0008117058985552997,
+          "avg_abnormal_signal_return_net": -0.00018829410144470112
+        },
+        "threshold_0.42": {
+          "num_events": 8,
+          "coverage": 0.005221932114882507,
+          "hit_ratio": 0.625,
+          "avg_signal_return": 0.007374357082881033,
+          "avg_signal_return_net": 0.0063743570828810325,
+          "median_signal_return": 0.010437145829200745,
+          "avg_abnormal_signal_return": 0.013620887020806549,
+          "avg_abnormal_signal_return_net": 0.012620887020806548
+        },
+        "threshold_0.45": {
+          "num_events": 0,
+          "coverage": 0.0,
+          "hit_ratio": null,
+          "avg_signal_return": null,
+          "avg_signal_return_net": null,
+          "median_signal_return": null,
+          "avg_abnormal_signal_return": null,
+          "avg_abnormal_signal_return_net": null
+        },
+        "threshold_0.50": {
+          "num_events": 0,
+          "coverage": 0.0,
+          "hit_ratio": null,
+          "avg_signal_return": null,
+          "avg_signal_return_net": null,
+          "median_signal_return": null,
+          "avg_abnormal_signal_return": null,
+          "avg_abnormal_signal_return_net": null
+        }
+      },
+      "mean_text_gate_weight": 0.5071590542793274,
+      "std_text_gate_weight": 0.0016348367789760232
+    }
+  },
+  "test": {
+    "sentiment": {
+      "accuracy": 0.009784735812133072,
+      "macro_precision": 0.0032615786040443573,
+      "macro_recall": 0.3333333333333333,
+      "macro_f1": 0.006459948320413436,
+      "confusion_matrix": [
+        [
+          0,
+          110,
+          0
+        ],
+        [
+          0,
+          15,
+          0
+        ],
+        [
+          0,
+          1408,
+          0
+        ]
+      ],
+      "predicted_class_counts": {
+        "bearish": 0,
+        "neutral": 1533,
+        "bullish": 0
+      },
+      "true_class_counts": {
+        "bearish": 110,
+        "neutral": 15,
+        "bullish": 1408
+      },
+      "expected_calibration_error": 0.8247767172447622,
+      "brier_score": 1.5716914553389048
+    },
+    "direction_1d": {
+      "accuracy": 0.3333333333333333,
+      "macro_precision": 0.2797514006239241,
+      "macro_recall": 0.30693224852151285,
+      "macro_f1": 0.28541099148120913,
+      "confusion_matrix": [
+        [
+          31,
+          217,
+          212
+        ],
+        [
+          54,
+          138,
+          244
+        ],
+        [
+          124,
+          171,
+          342
+        ]
+      ],
+      "predicted_class_counts": {
+        "bearish": 209,
+        "neutral": 526,
+        "bullish": 798
+      },
+      "true_class_counts": {
+        "bearish": 460,
+        "neutral": 436,
+        "bullish": 637
+      },
+      "expected_calibration_error": 0.03496117605077468,
+      "brier_score": 0.6723385830557309
+    },
+    "direction_5d": {
+      "accuracy": 0.45857795172863663,
+      "macro_precision": 0.28120269843193446,
+      "macro_recall": 0.33741824434028733,
+      "macro_f1": 0.30491929241175314,
+      "confusion_matrix": [
+        [
+          0,
+          123,
+          143
+        ],
+        [
+          0,
+          553,
+          311
+        ],
+        [
+          0,
+          253,
+          150
+        ]
+      ],
+      "predicted_class_counts": {
+        "bearish": 0,
+        "neutral": 929,
+        "bullish": 604
+      },
+      "true_class_counts": {
+        "bearish": 266,
+        "neutral": 864,
+        "bullish": 403
+      },
+      "expected_calibration_error": 0.024249568105288305,
+      "brier_score": 0.633983250156182
+    },
+    "volatility": {
+      "accuracy": 0.39921722113502933,
+      "macro_precision": 0.13307240704500978,
+      "macro_recall": 0.3333333333333333,
+      "macro_f1": 0.1902097902097902,
+      "confusion_matrix": [
+        [
+          0,
+          534,
+          0
+        ],
+        [
+          0,
+          612,
+          0
+        ],
+        [
+          0,
+          387,
+          0
+        ]
+      ],
+      "predicted_class_counts": {
+        "lower_vol": 0,
+        "normal_vol": 1533,
+        "higher_vol": 0
+      },
+      "true_class_counts": {
+        "lower_vol": 534,
+        "normal_vol": 612,
+        "higher_vol": 387
+      },
+      "expected_calibration_error": 0.1760233100613036,
+      "brier_score": 0.712886643106662
+    },
+    "event_backtest": {
+      "split": "test",
+      "transaction_cost_bps": 10.0,
+      "active_1d_fraction": 0.6568819308545336,
+      "hit_ratio_1d_active": 0.5789473684210527,
+      "avg_signal_return_1d_all": 0.0023507131510749455,
+      "avg_signal_return_1d_all_net": 0.0016938312202204116,
+      "avg_signal_return_1d_active": 0.0035785931088360393,
+      "avg_signal_return_1d_active_net": 0.002578593108836038,
+      "avg_abnormal_signal_return_1d_all": 0.0007162017627715418,
+      "avg_abnormal_signal_return_1d_all_net": 5.93198319170076e-05,
+      "avg_abnormal_signal_return_1d_active": 0.0010903051661656142,
+      "avg_abnormal_signal_return_1d_active_net": 9.030516616561321e-05,
+      "soft_signal_backtest_1d": {
+        "mean_abs_signal": 0.049029190093278885,
+        "active_fraction_abs_score_ge_0.10": 0.11545988258317025,
+        "hit_ratio_all_gross": 0.532941943900848,
+        "hit_ratio_active_gross": 0.4463276836158192,
+        "avg_signal_return_all": -4.1903094825102016e-05,
+        "avg_signal_return_all_net": -9.093229164136574e-05,
+        "avg_abnormal_signal_return_all": 6.6492548285168596e-06,
+        "avg_abnormal_signal_return_all_net": -4.237994289724156e-05
+      },
+      "active_5d_fraction": 0.39399869536855836,
+      "hit_ratio_5d_active": 0.3708609271523179,
+      "avg_signal_return_5d_all": -0.004762698527096634,
+      "avg_signal_return_5d_active": -0.012088107354369436,
+      "confidence_backtests_1d": {
+        "top_10pct_directional_confidence": {
+          "num_events": 153,
+          "mean_confidence": 0.40472039580345154,
+          "hit_ratio": 0.48366013071895425,
+          "avg_signal_return": -0.0017291086948829782,
+          "avg_signal_return_net": -0.002729108694882979,
+          "median_signal_return": -0.0011537353275343776,
+          "avg_abnormal_signal_return": 0.0023775517351216227,
+          "avg_abnormal_signal_return_net": 0.001377551735121622
+        },
+        "top_20pct_directional_confidence": {
+          "num_events": 307,
+          "mean_confidence": 0.39327916502952576,
+          "hit_ratio": 0.46579804560260585,
+          "avg_signal_return": -0.0018346647964980064,
+          "avg_signal_return_net": -0.0028346647964980075,
+          "median_signal_return": -0.001494303229264915,
+          "avg_abnormal_signal_return": 0.00013626769358741198,
+          "avg_abnormal_signal_return_net": -0.0008637323064125889
+        },
+        "top_30pct_directional_confidence": {
+          "num_events": 460,
+          "mean_confidence": 0.3858344554901123,
+          "hit_ratio": 0.5021739130434782,
+          "avg_signal_return": 0.000168503345380684,
+          "avg_signal_return_net": -0.0008314966546193169,
+          "median_signal_return": 0.00047751690726727247,
+          "avg_abnormal_signal_return": -0.0006967808645482443,
+          "avg_abnormal_signal_return_net": -0.0016967808645482452
+        },
+        "threshold_0.34": {
+          "num_events": 1108,
+          "coverage": 0.7227658186562296,
+          "hit_ratio": 0.5523465703971119,
+          "avg_signal_return": 0.002400788538323842,
+          "avg_signal_return_net": 0.001400788538323841,
+          "median_signal_return": 0.0029742957558482885,
+          "avg_abnormal_signal_return": 0.00043566977963306825,
+          "avg_abnormal_signal_return_net": -0.0005643302203669326
+        },
+        "threshold_0.36": {
+          "num_events": 636,
+          "coverage": 0.41487279843444225,
+          "hit_ratio": 0.5471698113207547,
+          "avg_signal_return": 0.001725182487273244,
+          "avg_signal_return_net": 0.000725182487273243,
+          "median_signal_return": 0.0027859483379870653,
+          "avg_abnormal_signal_return": -0.0001395135721791607,
+          "avg_abnormal_signal_return_net": -0.0011395135721791617
+        },
+        "threshold_0.38": {
+          "num_events": 247,
+          "coverage": 0.16112198303979125,
+          "hit_ratio": 0.44129554655870445,
+          "avg_signal_return": -0.0028902063559107334,
+          "avg_signal_return_net": -0.0038902063559107343,
+          "median_signal_return": -0.0022296553943306208,
+          "avg_abnormal_signal_return": 0.0008437850360328761,
+          "avg_abnormal_signal_return_net": -0.00015621496396712474
+        },
+        "threshold_0.40": {
+          "num_events": 85,
+          "coverage": 0.055446836268754074,
+          "hit_ratio": 0.43529411764705883,
+          "avg_signal_return": -0.0034284473146887168,
+          "avg_signal_return_net": -0.004428447314688717,
+          "median_signal_return": -0.0031941309571266174,
+          "avg_abnormal_signal_return": 0.0008126156158087884,
+          "avg_abnormal_signal_return_net": -0.00018738438419121242
+        },
+        "threshold_0.42": {
+          "num_events": 20,
+          "coverage": 0.01304631441617743,
+          "hit_ratio": 0.25,
+          "avg_signal_return": -0.008969856356270612,
+          "avg_signal_return_net": -0.009969856356270613,
+          "median_signal_return": -0.011959618888795376,
+          "avg_abnormal_signal_return": -0.0006033775032847188,
+          "avg_abnormal_signal_return_net": -0.0016033775032847197
+        },
+        "threshold_0.45": {
+          "num_events": 2,
+          "coverage": 0.001304631441617743,
+          "hit_ratio": 0.5,
+          "avg_signal_return": 0.0010717622935771942,
+          "avg_signal_return_net": 7.176229357719333e-05,
+          "median_signal_return": 0.0010717622935771942,
+          "avg_abnormal_signal_return": 0.01302456425037235,
+          "avg_abnormal_signal_return_net": 0.01202456425037235
+        },
+        "threshold_0.50": {
+          "num_events": 0,
+          "coverage": 0.0,
+          "hit_ratio": null,
+          "avg_signal_return": null,
+          "avg_signal_return_net": null,
+          "median_signal_return": null,
+          "avg_abnormal_signal_return": null,
+          "avg_abnormal_signal_return_net": null
+        }
+      },
+      "mean_text_gate_weight": 0.506599485874176,
+      "std_text_gate_weight": 0.0016385371563956141
+    }
+  },
+  "data": {
+    "num_rows": 13930,
+    "train_rows": 10865,
+    "validation_rows": 1532,
+    "test_rows": 1533,
+    "symbols": [
+      "AAPL",
+      "ABNB",
+      "ADBE",
+      "ADI",
+      "ADP",
+      "AMAT",
+      "AMD",
+      "AMZN",
+      "ARKK",
+      "ASML",
+      "AVGO",
+      "BA",
+      "BABA",
+      "BAC",
+      "BIDU",
+      "BMY",
+      "CAT",
+      "CMCSA",
+      "COIN",
+      "COST",
+      "CRM",
+      "CSCO",
+      "CVX",
+      "DHR",
+      "DIS",
+      "ELF",
+      "GE",
+      "GM",
+      "GOOGL",
+      "HD",
+      "IBM",
+      "INTC",
+      "IOVA",
+      "JNJ",
+      "JPM",
+      "KO",
+      "LULU",
+      "MA",
+      "MCD",
+      "MDB",
+      "META",
+      "MRNA",
+      "MSFT",
+      "MU",
+      "NFLX",
+      "NKE",
+      "NVDA",
+      "ORCL",
+      "PEP",
+      "PFE",
+      "PLTR",
+      "PYPL",
+      "QCOM",
+      "RCL",
+      "SHOP",
+      "SNOW",
+      "SPOT",
+      "TGT",
+      "TSLA",
+      "TSM",
+      "TXN",
+      "UAL",
+      "UBER",
+      "UNH",
+      "UPS",
+      "WMT",
+      "XOM",
+      "ZM"
+    ],
+    "date_min": "2022-08-12 21:13:30+00:00",
+    "date_max": "2023-10-04 20:50:09+00:00",
+    "volatility_thresholds": {
+      "q33": 0.01998170108283479,
+      "q66": 0.049223492463942975
+    }
+  }
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3f6c77bf9ca6e4ab1c9db29f57173379f7e8daeedbdf439271456979730986b
+size 869703256

modeling_finimpact.py

@@ -0,0 +1,6 @@
+"""
+Minimal loading helper for the FinImpact gated fusion checkpoint.
+
+Use this repository with the original training script for the full class definition.
+The uploaded model is a custom torch.nn.Module, not a vanilla AutoModelForSequenceClassification.
+"""

tokenizer_config.json

@@ -0,0 +1,24 @@
+{
+  "add_prefix_space": true,
+  "backend": "tokenizers",
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "eos_token": "[SEP]",
+  "extra_special_tokens": [
+    "[PAD]",
+    "[CLS]",
+    "[SEP]"
+  ],
+  "is_local": false,
+  "local_files_only": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "split_by_punct": false,
+  "tokenizer_class": "DebertaV2Tokenizer",
+  "unk_id": 3,
+  "unk_token": "[UNK]",
+  "vocab_type": "spm"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3bf5789497aeb9dde5f757914e12ec101b234de047a1d0df789825a513676b4
+size 5265

training_config.json

@@ -0,0 +1,127 @@
+{
+  "model_repo_id": "kyLELEng/finimpact-direction1d-v4",
+  "encoder_name": "microsoft/deberta-v3-large",
+  "news_dataset_id": "NickyNicky/finance-financialmodelingprep-stock-news-sentiments-rss-feed",
+  "news_split": "train",
+  "price_dataset_id": "siddharthmb/stocks-ohlcv",
+  "price_dataset_file": "ohlcv.csv",
+  "tickers": [
+    "AAPL",
+    "MSFT",
+    "AMZN",
+    "GOOGL",
+    "NVDA",
+    "TSLA",
+    "AMD",
+    "INTC",
+    "JPM",
+    "BAC",
+    "V",
+    "MA",
+    "XOM",
+    "META",
+    "NFLX",
+    "UAL",
+    "HD",
+    "BA",
+    "PYPL",
+    "CRM",
+    "DIS",
+    "NKE",
+    "WMT",
+    "COST",
+    "PFE",
+    "MRNA",
+    "CVX",
+    "KO",
+    "PEP",
+    "CSCO",
+    "ABNB",
+    "ADBE",
+    "ADI",
+    "ADP",
+    "AMAT",
+    "ARKK",
+    "ASML",
+    "AVGO",
+    "BABA",
+    "BIDU",
+    "BMY",
+    "CAT",
+    "CMCSA",
+    "COIN",
+    "DHR",
+    "ELF",
+    "GE",
+    "GEV",
+    "GM",
+    "IBM",
+    "IOVA",
+    "JNJ",
+    "LULU",
+    "MCD",
+    "MDB",
+    "MU",
+    "ORCL",
+    "PLTR",
+    "QCOM",
+    "RCL",
+    "SHOP",
+    "SNOW",
+    "SPOT",
+    "TGT",
+    "TSM",
+    "TXN",
+    "UBER",
+    "UNH",
+    "UPS",
+    "ZM"
+  ],
+  "benchmark_symbols": [
+    "SPY",
+    "QQQ",
+    "DIA"
+  ],
+  "beta_window": 60,
+  "max_news_rows": 250000,
+  "max_examples": 100000,
+  "min_text_chars": 20,
+  "max_length": 256,
+  "train_fraction": 0.78,
+  "validation_fraction": 0.11,
+  "direction_threshold_vol_mult_1d": 0.25,
+  "direction_threshold_vol_mult_5d": 0.5,
+  "min_direction_threshold_1d": 0.0015,
+  "min_direction_threshold_5d": 0.003,
+  "fusion_dim": 256,
+  "dropout": 0.2,
+  "freeze_encoder": false,
+  "sentiment_loss_weight": 0.0,
+  "direction_1d_loss_weight": 1.0,
+  "direction_5d_loss_weight": 0.0,
+  "volatility_loss_weight": 0.0,
+  "class_weight_cap": 2.5,
+  "focal_loss_gamma": 1.5,
+  "return_alignment_loss_weight": 0.1,
+  "return_confidence_loss_weight": 0.03,
+  "return_scale": 0.02,
+  "transaction_cost_bps": 10.0,
+  "per_device_train_batch_size": 4,
+  "per_device_eval_batch_size": 32,
+  "gradient_accumulation_steps": 1,
+  "learning_rate": 2e-06,
+  "weight_decay": 0.01,
+  "warmup_ratio": 0.06,
+  "max_grad_norm": 0.5,
+  "num_train_epochs": 3.0,
+  "max_steps": -1,
+  "logging_steps": 50,
+  "eval_strategy": "epoch",
+  "save_strategy": "no",
+  "bf16": false,
+  "fp16": false,
+  "random_seed": 13,
+  "csv_chunksize": 500000,
+  "output_dir": "finimpact-direction1d-v4",
+  "push_to_hub": true
+}