Files
foxhunt/optimize_batch_sizes.sh
jgrusewski 35feadf55e 🚀 Wave 160 Phase 6: CUDA Mandatory + TDD Testing + TFT Complete (21 Agents)
## Major Achievements

### 1. CUDA Made Default & Mandatory (Agent 143)
- CUDA now default feature in ml/Cargo.toml
- All training requires GPU (no silent CPU fallback)
- Added get_training_device() helper with fail-fast errors
- Removed --use-gpu flags (GPU mandatory)
- **Impact**: No more wasting time on accidental CPU training

### 2. TFT Training COMPLETE (Agent 144)
-  Training completed successfully in 7.6 minutes
-  Early stopping at epoch 100/200 (best val loss: 0.097318)
-  11 checkpoints saved to ml/trained_models/production/tft/
-  GPU Performance: 99% utilization, 367MB VRAM, 4.4s/epoch
-  10x speedup vs CPU (4.4s vs 43-55s per epoch)
- **Status**: PRODUCTION READY

### 3. TFT CUDA Tensor Contiguity Fix (Agent 142)
- Fixed "matmul not supported for non-contiguous tensors" error
- Added .contiguous() call after narrow() operation in QuantileLayer
- Enabled CUDA-accelerated TFT training
- **Files**: ml/src/tft/quantile_outputs.rs

### 4. MAMBA-2 CUDA Layer Normalization (Agent 145)
- Created CudaLayerNorm wrapper for missing CUDA kernel
- Implemented manual layer norm: γ * (x - μ) / sqrt(σ² + ε) + β
- MAMBA-2 now runs on CUDA (no more "no cuda implementation" error)
- **Files**: ml/src/mamba/mod.rs

### 5. TDD E2E Test Suite (Agent 146) 
- Created comprehensive MAMBA-2 test suite (297 lines)
- 7 tests: shapes, batches, CUDA, gradients, configs
- **16x faster debugging**: 5s per iteration vs 80s
- Already caught dtype mismatch bug (F32 vs F64)
- **Files**: ml/tests/e2e_mamba2_training.rs

## Agent Summary (Agents 126-146)

### Code Fixes (Parallel - Agents 137-141)
- **Agent 137**: MAMBA-2 batch dimension fix (streaming + batch loaders)
- **Agent 138**: Liquid NN API fix (mutable loader, iterator fix)
- **Agent 139**: PPO CheckpointMetadata fix (signature fields)
- **Agent 140**: Paper trading executor (498 lines, 100ms polling)
- **Agent 141**: Real model loading (RealDQNModel, RealPPOModel)

### Infrastructure (Agents 143-146)
- **Agent 143**: CUDA mandatory (Cargo.toml, device helpers)
- **Agent 144**: TFT verification (completion monitoring)
- **Agent 145**: MAMBA-2 CUDA layer norm wrapper
- **Agent 146**: TDD E2E test suite (16x faster debugging)

## Files Modified

### Core ML Infrastructure
- ml/Cargo.toml: Added default = ["minimal-inference", "cuda"]
- ml/src/lib.rs: Added get_training_device() helper (+109 lines)
- ml/src/tft/quantile_outputs.rs: Fixed tensor contiguity
- ml/src/mamba/mod.rs: Added CudaLayerNorm wrapper (+41 lines)

### Training Scripts
- ml/examples/train_tft_dbn.rs: Removed --use-gpu flag
- ml/examples/train_ppo.rs: Removed --use-gpu flag
- ml/examples/train_mamba2_dbn.rs: Forced CUDA-only mode
- ml/examples/train_liquid_dbn.rs: Fixed API usage

### Data Loaders
- ml/src/data_loaders/dbn_sequence_loader.rs: Fixed batch dimensions
- ml/src/data_loaders/streaming_dbn_loader.rs: Fixed batch dimensions

### Trading Service
- services/trading_service/src/paper_trading_executor.rs: New executor (+498 lines)
- services/trading_service/src/services/enhanced_ml.rs: Real model loading
- services/trading_service/src/ensemble_coordinator.rs: Integration

### Tests
- ml/tests/e2e_mamba2_training.rs: New TDD test suite (+297 lines)

### Trainers
- ml/src/trainers/tft.rs: Fixed CheckpointMetadata signature fields

## Performance Metrics

### TFT Training
- Duration: 7.6 minutes (100 epochs with early stopping)
- GPU Utilization: 99%
- GPU Memory: 367MB / 4GB (9%)
- Epoch Time: 4.4 seconds (vs 43-55s on CPU)
- Speedup: 10x vs CPU
- Status:  PRODUCTION READY

### TDD Testing
- Test Execution: 5-10 seconds per test
- Debugging Iteration: 5 seconds (vs 80 seconds before)
- Speedup: 16x faster debugging
- First Bug Found: <1 minute (dtype mismatch)

## Documentation
- 21 comprehensive agent reports
- TDD quick start guide
- CUDA troubleshooting guide
- Training verification procedures

## Next Steps
1. Fix MAMBA-2 dtype mismatch (F32→F64) - 2 minutes
2. Run MAMBA-2 tests until passing - 5-10 minutes
3. Launch full MAMBA-2 training - 200 epochs
4. Launch Liquid NN training

## System Status
- TFT:  COMPLETE (production ready)
- MAMBA-2: 🧪 IN TESTING (TDD suite ready)
- CUDA:  DEFAULT (mandatory for training)
- Tests:  16x faster debugging

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-10-14 23:13:34 +02:00

102 lines
2.7 KiB
Bash
Executable File

#!/bin/bash
# GPU Batch Size Optimization Script for RTX 3050 Ti (4GB VRAM)
#
# Tests optimal batch sizes for TFT, MAMBA-2, and Liquid models
# Generates BATCH_SIZE_OPTIMIZATION_REPORT.md with recommendations
set -e
echo "==================================="
echo "GPU Batch Size Optimization"
echo "==================================="
echo ""
# Check if nvidia-smi is available
if ! command -v nvidia-smi &> /dev/null; then
echo "ERROR: nvidia-smi not found. This script requires NVIDIA GPU."
exit 1
fi
# Display GPU info
echo "GPU Information:"
nvidia-smi --query-gpu=name,memory.total,driver_version --format=csv,noheader
echo ""
# Check CUDA availability
echo "Checking CUDA setup..."
if [ -d "/usr/local/cuda" ]; then
echo "✓ CUDA found at /usr/local/cuda"
nvcc --version | head -n 1
else
echo "⚠ CUDA not found at /usr/local/cuda (will fall back to CPU)"
fi
echo ""
# Build the optimization tool in release mode
echo "Building optimization tool (release mode)..."
cargo build -p ml --example optimize_batch_sizes --release
if [ $? -ne 0 ]; then
echo "ERROR: Failed to build optimization tool"
exit 1
fi
echo "✓ Build complete"
echo ""
# Run the optimization
echo "Running batch size optimization..."
echo "This will test:"
echo " - TFT: batch sizes [16, 32, 64, 128]"
echo " - MAMBA-2: batch sizes [8, 16, 32]"
echo " - Liquid: batch sizes [16, 32, 64]"
echo ""
echo "Estimated runtime: 3-5 minutes"
echo ""
# Monitor VRAM usage in background
echo "Starting VRAM monitor..."
(
while true; do
nvidia-smi --query-gpu=memory.used,memory.total --format=csv,noheader,nounits | \
awk '{printf "VRAM: %d MB / %d MB (%.1f%%)\r", $1, $2, ($1/$2)*100}'
sleep 1
done
) &
MONITOR_PID=$!
# Run the optimization
cargo run -p ml --example optimize_batch_sizes --release
# Kill the monitor
kill $MONITOR_PID 2>/dev/null || true
echo ""
# Check if report was generated
if [ -f "BATCH_SIZE_OPTIMIZATION_REPORT.md" ]; then
echo ""
echo "==================================="
echo "Optimization Complete!"
echo "==================================="
echo ""
echo "Report generated: BATCH_SIZE_OPTIMIZATION_REPORT.md"
echo ""
# Extract recommendations
echo "Recommended Batch Sizes:"
grep -A 10 "## Optimization Summary" BATCH_SIZE_OPTIMIZATION_REPORT.md | \
grep -E "^\| (TFT|MAMBA-2|Liquid)" | \
awk -F'|' '{print " " $2 " -> batch_size = " $3}' | \
sed 's/ //' | sed 's/^ *//'
echo ""
echo "Next Steps:"
echo "1. Review BATCH_SIZE_OPTIMIZATION_REPORT.md for detailed results"
echo "2. Update model configurations with recommended batch sizes"
echo "3. Test training with optimized batch sizes"
else
echo ""
echo "ERROR: Report not generated"
exit 1
fi