maverick90024 commited on
Commit
b86536b
·
verified ·
1 Parent(s): d65d4b0

Export ensemble models for BTC-USD (1h)

Browse files
balanced_lookback60_6h/README.md ADDED
@@ -0,0 +1,98 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - time-series
4
+ - forecasting
5
+ - temporal
6
+ - trading
7
+ - btc-usd
8
+ - ensemble
9
+ library_name: temporal-forecasting
10
+ ---
11
+
12
+ # Temporal Trading Model: BTC-USD (Ensemble Member 5/7: balanced (lookback=60))
13
+
14
+ This is a pre-trained Temporal transformer model for time series forecasting of BTC-USD. This model is part of a consensus ensemble.
15
+
16
+ ## Ensemble Context
17
+
18
+ This model is **member 5 of 7** in a consensus ensemble for BTC-USD (1h).
19
+
20
+ The ensemble uses 7 models with different lookback periods and focus strategies to generate diverse forecasts. These forecasts are then aggregated using multiple consensus strategies (gradient, confidence, timeframe, volatility, mean reversion, acceleration, swing, risk-adjusted) to produce robust trading signals.
21
+
22
+ ### All Ensemble Members
23
+
24
+ 1. [momentum (lookback=30)](../momentum_lookback30_12h)
25
+ 2. [momentum (lookback=30)](../momentum_lookback30_6h)
26
+ 3. [balanced (lookback=45)](../balanced_lookback45_12h)
27
+ 4. [balanced (lookback=60)](../balanced_lookback60_24h)
28
+ 5. **balanced** (lookback=60) - *This model*
29
+ 6. [balanced (lookback=60)](../balanced_lookback60_12h)
30
+ 7. [mean_reversion (lookback=45)](../mean_reversion_lookback45_12h)
31
+
32
+ ### How the Ensemble Works
33
+
34
+ 1. Each member model generates independent forecasts using its specific lookback window and focus
35
+ 2. Multiple consensus strategies analyze the ensemble's forecasts from different perspectives
36
+ 3. Each strategy produces action recommendations (BUY/SELL/HOLD) with confidence scores
37
+ 4. Final consensus aggregates all strategy recommendations into a unified trading signal
38
+
39
+
40
+ ## Model Details
41
+
42
+ - **Symbol**: BTC-USD
43
+ - **Interval**: 1h
44
+ - **Lookback Window**: 60 periods
45
+ - **Forecast Horizon**: 6 periods
46
+ - **Focus**: balanced
47
+ - **Training Date**: 2025-11-10 08:02 UTC
48
+ - **Training Epochs**: 10
49
+ - **Best Validation Loss**: 0.20215057868223923
50
+
51
+ ## Training Dataset
52
+
53
+ - **Date Range**: 2020-11-12T00:49:00 to 2025-11-10
54
+ - **Training Samples**: 1837645
55
+ - **Lookback Period**: 60 1h intervals
56
+
57
+ ## Model Architecture
58
+
59
+ - **Model Type**: Temporal Transformer
60
+ - **d_model**: 512
61
+ - **Encoder Layers**: None
62
+ - **Decoder Layers**: None
63
+
64
+ ## Features
65
+
66
+ Input features used: 17 features
67
+
68
+ ## Usage
69
+
70
+ ```python
71
+ from strategies.model_cache import get_model_cache
72
+
73
+ # Import from HuggingFace
74
+ cache = get_model_cache()
75
+ model_path, scaler_path, metadata = cache.import_from_huggingface(
76
+ repo_id="YOUR_REPO_ID",
77
+ symbol="BTC-USD",
78
+ interval="1h",
79
+ lookback=60,
80
+ focus="balanced",
81
+ forecast_horizon=6
82
+ )
83
+ ```
84
+
85
+ ## License
86
+
87
+ GPL-3.0-or-later
88
+
89
+ ## Citation
90
+
91
+ ```
92
+ @software{temporal_trading_model,
93
+ title = {Temporal Trading Model: BTC-USD (balanced)},
94
+ author = {Unidatum Integrated Products LLC},
95
+ year = {2025},
96
+ url = {https://github.com/OptimalMatch/temporal-trading-agents}
97
+ }
98
+ ```
balanced_lookback60_6h/config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cache_key": "BTC-USD_1h_balanced_efd785a6dc29",
3
+ "symbol": "BTC-USD",
4
+ "interval": "1h",
5
+ "lookback": 60,
6
+ "focus": "balanced",
7
+ "forecast_horizon": 6,
8
+ "feature_columns": [
9
+ "Close",
10
+ "Open",
11
+ "High",
12
+ "Low",
13
+ "Volume",
14
+ "Returns",
15
+ "Log_Returns",
16
+ "MA_7",
17
+ "MA_21",
18
+ "MA_50",
19
+ "Volatility_7",
20
+ "Volatility_21",
21
+ "RSI",
22
+ "Momentum_7",
23
+ "Momentum_21",
24
+ "Volume_Ratio",
25
+ "Price_Range"
26
+ ],
27
+ "training_timestamp": "2025-11-10T08:02:37.669175",
28
+ "data_end_timestamp": "2025-11-10T05:00:00",
29
+ "training_epochs": 10,
30
+ "best_val_loss": 0.20215057868223923,
31
+ "model_architecture": {
32
+ "d_model": 512,
33
+ "num_encoder_layers": null,
34
+ "num_decoder_layers": null
35
+ },
36
+ "data_start_timestamp": "2020-11-12T00:49:00",
37
+ "training_samples": 1837645
38
+ }
balanced_lookback60_6h/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6379b19be54e953acf7d362821a1ef1908ccd4dcea6f61c0e2cbe44e9f791149
3
+ size 186982282
balanced_lookback60_6h/scaler.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:636f7b1069b65cfd929c3bc14a3b68a74ae79eb1965e3c67707eed742ad5ebe2
3
+ size 857