12fdd18223
Deleted: - DQN::compute_loss_internal (280 lines) — old Candle forward+loss - DQN::train_step (55 lines) — old Candle training step - DQN::compute_gradients (47 lines) — old gradient accumulation - ComputeLossResult struct — only used by deleted functions - RegimeConditionalDQN::train_step (65 lines) — old dispatch - RegimeConditionalDQN::train_step_gpu_regime (100 lines) — old GPU path - RegimeConditionalDQN::compute_gradients_gpu (130 lines) — old regime gradients - RegimeConditionalDQN::compute_gradients (92 lines) — old dispatch - DQNAgentType::train_step dispatch — dead - DQNAgentType::compute_gradients dispatch — dead - GpuDqnTrainer::upload_batch (71 lines) — old CPU→GPU upload - train_step.rs (500 lines) — entire module including ensure_fused_ctx - dqn_benchmark.rs — used old train_step - examples.rs — used old train_step - validation/adapters.rs (289 lines) — used old train_step - dqn/trainable_adapter.rs — used old train_step - gpu_smoketest.rs — tested old train_step - Gradient accumulation path in training_loop.rs (144 lines) - IQN d_h_s2().clone() → raw pointer (zero alloc) - Causal intervention format! string alloc removed - Dead HER relabel functions (320 lines) Kept: - ensure_fused_ctx logic inlined into training_loop.rs - set_noise_sigma_scale re-added to RegimeConditionalDQN Fixed: - GpuReplayBuffer max_batch_size wired from batch_size parameter (was hardcoded 1024, blocking batch_size=8192) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
3.5 KiB
3.5 KiB
name, description
| name | description |
|---|---|
| claude-flow-help | Show Claude-Flow commands and usage |
Claude-Flow Commands
🌊 Claude-Flow: Agent Orchestration Platform
Claude-Flow is the ultimate multi-terminal orchestration platform that revolutionizes how you work with Claude Code.
Core Commands
🚀 System Management
./claude-flow start- Start orchestration system./claude-flow start --ui- Start with interactive process management UI./claude-flow status- Check system status./claude-flow monitor- Real-time monitoring./claude-flow stop- Stop orchestration
🤖 Agent Management
./claude-flow agent spawn <type>- Create new agent./claude-flow agent list- List active agents./claude-flow agent info <id>- Agent details./claude-flow agent terminate <id>- Stop agent
📋 Task Management
./claude-flow task create <type> "description"- Create task./claude-flow task list- List all tasks./claude-flow task status <id>- Task status./claude-flow task cancel <id>- Cancel task./claude-flow task workflow <file>- Execute workflow
🧠 Memory Operations
./claude-flow memory store "key" "value"- Store data./claude-flow memory query "search"- Search memory./claude-flow memory stats- Memory statistics./claude-flow memory export <file>- Export memory./claude-flow memory import <file>- Import memory
⚡ SPARC Development
./claude-flow sparc "task"- Run SPARC orchestrator./claude-flow sparc modes- List all 17+ SPARC modes./claude-flow sparc run <mode> "task"- Run specific mode./claude-flow sparc tdd "feature"- TDD workflow./claude-flow sparc info <mode>- Mode details
🐝 Swarm Coordination
./claude-flow swarm "task" --strategy <type>- Start swarm./claude-flow swarm "task" --background- Long-running swarm./claude-flow swarm "task" --monitor- With monitoring./claude-flow swarm "task" --ui- Interactive UI./claude-flow swarm "task" --distributed- Distributed coordination
🌍 MCP Integration
./claude-flow mcp status- MCP server status./claude-flow mcp tools- List available tools./claude-flow mcp config- Show configuration./claude-flow mcp logs- View MCP logs
🤖 Claude Integration
./claude-flow claude spawn "task"- Spawn Claude with enhanced guidance./claude-flow claude batch <file>- Execute workflow configuration
🌟 Quick Examples
Initialize with SPARC:
npx -y claude-flow@latest init --sparc
Start a development swarm:
./claude-flow swarm "Build REST API" --strategy development --monitor --review
Run TDD workflow:
./claude-flow sparc tdd "user authentication"
Store project context:
./claude-flow memory store "project_requirements" "e-commerce platform specs" --namespace project
Spawn specialized agents:
./claude-flow agent spawn researcher --name "Senior Researcher" --priority 8
./claude-flow agent spawn developer --name "Lead Developer" --priority 9
🎯 Best Practices
- Use
./claude-flowinstead ofnpx claude-flowafter initialization - Store important context in memory for cross-session persistence
- Use swarm mode for complex tasks requiring multiple agents
- Enable monitoring for real-time progress tracking
- Use background mode for tasks > 30 minutes