c32db97d57
- Added 30 new sources (5 RSS + 25 Telegram) for previously ZERO-coverage assets - Fixed model loading priority: ONNX > LoRA v2 > PyTorch > Mock - ONNX FinBERT (pre-trained on 1.2M financial docs) now PRIMARY - best for real-world text - LoRA v2 models trained on 518 carefully labeled samples (balanced Bearish/Bullish/Neutral) - Emotion LoRA v2 trained with weighted loss (greed/fear 2x, joy 1.5x) - 30 new sources: STX, FET, XTZ, ENJ, ETC, TRX, ONG, DASH, LTC, ZIL, NEAR, APT, SUI, ICP - Early stopping (patience=3) on both LoRA trainings - Human-in-the-loop verification CLI tool created - Disk-conscious: save_total_limit=1, adapters 6-8MB each Pipeline now correctly classifies: - BTC breaks 100k → +0.54 Bullish ✅ - Major hack → -0.23 Bearish ✅ - HODL → +0.91 Bullish ✅ - Rug pull → -0.30 Bearish ✅ - SEC sues → -0.30 Bearish ✅ - ETF approval → +0.32 Bullish ✅ - Whale accumulation → +0.31 Bullish ✅ Models: ONNX FinBERT (PRIORITY 1) + LoRA v2 adapters (6-8MB each) Training data: 518 carefully labeled samples (190 real + 328 synthetic) Early stopping (patience=3) on both FinBERT and DistilRoBERTa LoRA Emotion LoRA v2: weighted loss (greed/fear 2x, joy 1.5x) + early stopping
68 lines
1.8 KiB
Python
68 lines
1.8 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
Combine all labeled datasets and retrain LoRA models.
|
|
"""
|
|
|
|
import json
|
|
import random
|
|
from pathlib import Path
|
|
|
|
# Load all labeled data
|
|
all_data = []
|
|
|
|
for fname in [
|
|
"labeled_verified.jsonl",
|
|
"labeled_expanded.jsonl",
|
|
"labeled_large.jsonl",
|
|
"labeled_output.jsonl",
|
|
"labeled_real_world.jsonl",
|
|
]:
|
|
path = Path(f"data/{fname}")
|
|
if path.exists():
|
|
with open(path) as f:
|
|
for line in f:
|
|
try:
|
|
item = json.loads(line.strip())
|
|
# Normalize format
|
|
labels = item.get("labels", {})
|
|
text = item.get("text", item.get("raw_text", ""))
|
|
if text and labels.get("sentiment"):
|
|
all_data.append({
|
|
"text": text,
|
|
"sentiment": labels["sentiment"],
|
|
"event_type": labels.get("event_type", "unknown"),
|
|
"entities": labels.get("entities", []),
|
|
})
|
|
except Exception as e:
|
|
print(f"Error in {fname}: {e}")
|
|
|
|
# Deduplicate by text hash
|
|
seen = set()
|
|
unique = []
|
|
for d in all_data:
|
|
h = hash(d["text"][:200])
|
|
if h not in seen:
|
|
seen.add(h)
|
|
unique.append(d)
|
|
|
|
print(f"Total unique samples: {len(unique)}")
|
|
|
|
# Split
|
|
random.shuffle(unique)
|
|
split = int(0.9 * len(unique))
|
|
train_data = unique[:split]
|
|
val_data = unique[split:]
|
|
|
|
print(f"Train: {len(train_data)} | Val: {len(val_data)}")
|
|
|
|
# Save combined dataset
|
|
with open("data/combined_train.jsonl", "w") as f:
|
|
for d in train_data:
|
|
f.write(json.dumps(d) + "\n")
|
|
|
|
with open("data/combined_val.jsonl", "w") as f:
|
|
for d in val_data:
|
|
f.write(json.dumps(d) + "\n")
|
|
|
|
print("Saved combined datasets")
|