'checkpoint' 검색 결과

Guide for checkpointing — saving, loading, and resuming training with CheckpointRecord. Use when the user asks about saving weights, resuming training, checkpoint management, or the checkpoint lifecycle.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Plan and troubleshoot large language and vision model training and inference. This guide uses hardware measurements to identify bottlenecks, address memory shortages and unstable training, and improve checkpointing, recovery, and performance across multiple GPUs and machines.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Diagnosing slow training and GPU memory shortages
• Investigating loss spikes and invalid numerical values
• Debugging communication hangs across GPUs and machines

Save the current code state in a local Git commit before risky changes, then return to that checkpoint if needed. Useful for multi-file refactoring and trying another approach after an unsuccessful edit.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Saving a checkpoint before multi-file refactoring
• Returning to saved code after unsuccessful changes
• Cleaning up temporary commits before pushing a PR branch

Configure a terminal status display for Diverga research projects. Show the project name, research stage, checkpoint progress, and memory health, choose among four display presets, and set up or repair the Claude Code integration.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Monitor research progress in the terminal
• Show detailed checkpoint completion status
• Focus the display on memory and context health

Restore project files from a previous Git checkpoint after previewing affected files and receiving explicit confirmation. Create a safety tag before restoration, preserve recorded lessons, and keep the current branch and commit history in place.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Listing available project checkpoints
• Restoring files after an unsuccessful change
• Previewing files affected by a rollback

Save a structured state snapshot before stopping or handing off. Captures the active plan, recent decisions, file pointers (with line numbers), open questions, and the next 1–3 actions into a checkpoint file under `quality_reports/checkpoints/`. Optionally proposes `[LEARN]` entries to add to MEMORY.md. Use when user says "checkpoint", "save state", "snapshot before I stop", "where am I", "wrap up the session for handoff", or before a long break / model switch / collaborator handoff. Companion to (NOT replacement for) the narrative session-log workflow.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Debug validator checkpoint inconsistencies where some validators are behind others. Use when alerts mention "checkpoint inconsistency", "validators behind", or "inconsistent latest checkpoints", or when asked to debug validator sets, investigate validator delays, or troubleshoot metadata fetch failures for a chain. Defaults to default_ism app context if not specified.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Export an approved fine-tuned model into a format suited to its hardware and runtime. Choose weight packaging and quantization, which reduces numerical precision, then compare outputs before and after export to check for regressions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Exporting a checkpoint approved for deployment
• Choosing a model format for a target GPU or laptop
• Comparing outputs before and after export

Evaluate whether a fine-tuned model is ready for release by checking training data, general capability loss, comparisons with the base model, and staged production rollout. Produces an evidence-based promotion or rejection report with one prioritized remedy when rejected.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Deciding whether to release a fine-tuned model
• Checking whether training reduced general capabilities
• Reassessing a model against updated reference evaluations

Build AI agents with LangGraph by connecting processing steps into a graph. Define shared state, conditional paths, tool calls, saved checkpoints, and human review points to create workflows that retain information as they run.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Building an agent that calls tools in a loop
• Sharing accumulated findings between multiple agents
• Routing requests to different processing paths

Configure large language model pretraining with torchtitan across multiple GPUs. Covers ways to split training work, launch jobs on computing clusters, enable Float8 numerical processing, and save or resume training through distributed checkpoints.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Pretraining Llama 3.1 8B on eight GPUs
• Configuring multi-node training with SLURM
• Enabling Float8 training on H100 GPUs

Turn a repeated manual task into a reusable skill you can call by name or schedule. It separates changing inputs from fixed steps, adds approval checkpoints, and tests the workflow against a known result before registration and team sharing.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Turning a weekly routine into a reusable skill
• Capturing a workflow demonstrated in conversation
• Adding approval checkpoints to business tasks

Check that changes to existing PEFT methods still work with saved training checkpoints. The skill guides compatibility tests, configuration checks, benchmark comparisons, and reporting, including comparisons of model outputs before and after a change.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Checking compatibility after a forward method rewrite
• Changing state_dict structure while supporting old checkpoints
• Reviewing defaults for new configuration fields

Helps choose and tune Cloud Storage FUSE mounts, which expose bucket objects through file paths. It covers cache sizing, training performance, checkpoint write safety, and cost diagnosis on GKE, Compute Engine, and Cloud Run.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Access training data through file paths
• Choose between FUSE, native reads, and shared file systems
• Check checkpoint and concurrent write safety

Guides adapting OpenPI robot models to custom datasets and serving them to produce actions from observations. Covers JAX and PyTorch training, checkpoint conversion, and troubleshooting for normalization statistics and GPU memory.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Fine-tune OpenPI models on custom robot datasets
• Convert JAX checkpoints to PyTorch
• Serve policies for ALOHA, DROID, or LIBERO

Add PyTorch FSDP2 to training scripts so model state can be split across GPUs. Covers distributed initialization, mixed precision, CPU offloading, and checkpoint saving and loading for models that exceed a single GPU's memory.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Training a model that exceeds one GPU's memory
• Adding FSDP2 to an existing training script
• Configuring mixed precision and CPU offloading

Configure large language model pretraining with torchtitan across multiple GPUs. Covers ways to split training work, launch jobs on computing clusters, enable Float8 numerical processing, and save or resume training through distributed checkpoints.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Pretraining Llama 3.1 8B on eight GPUs
• Configuring multi-node training with SLURM
• Enabling Float8 training on H100 GPUs

Design OpenRig watchdog policies for stalled or drifting agent work. Choose a small wake-up message, a refocus, or a broader alignment check, and separate monitoring frequency from notification frequency using evidence about owners and artifacts.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing an intervention for an inactive task owner
• Drafting messages that correct workflow drift
• Separating monitoring and notification intervals

Estimate GPU memory, training step time, checkpoint costs, and recovery overhead for training or reinforcement learning. Compare state partitioning, recomputation, offloading, and device counts to assess whether a job can finish within its deadline.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Check whether training can finish by a deadline
• Compare memory-saving methods and their time costs
• Plan checkpoint intervals and recovery overhead

Keep research progress, decisions, and pending checkpoints across sessions. Store project context locally, recover essential details when returning, and track why decisions changed so research work can resume with a clear next step.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Resuming research after a break
• Reviewing decisions and their rationale
• Checking pending research milestones

Continue KERMT pretraining on a custom SMILES corpus with a grover_base, cmim, or hybrid checkpoint. Use a local checkpoint or optionally download a pinned Hugging Face model bundle using HF_TOKEN if configured. Run containerized training and write model bundles, prepared data, logs, and checkpoints to user-selected host directories.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Save bounded progress as kernel.checkpoint/v1 for exact resume after reset. Triggers: checkpoint, save progress, context reset, compact soon, long task.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Expert blueprint for player mortality loops: progress-index checkpoints, soul graves, ghost/spirit layers, I-frame restitution, world-progress persistence, and death analytics. Use when implementing respawn, shrine checkpoints, corpse-run retrieval, or second-chance mechanics. Keywords revival, respawn, checkpoint, soul grave, ghost mode, I-frames, death analytics, progress index.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Expert blueprint for player mortality loops: progress-index checkpoints, soul graves, ghost/spirit layers, I-frame restitution, world-progress persistence, and death analytics. Use when implementing respawn, shrine checkpoints, corpse-run retrieval, or second-chance mechanics. Keywords revival, respawn, checkpoint, soul grave, ghost mode, I-frames, death analytics, progress index.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.