Loading...
Loading...
Found 100 Skills
Accepts Triton operator implementations, automatically invokes Torch small operator implementations (CPU or NPU) for precision comparison, and generates precision reports. It is used when users need to verify the correctness and precision of Triton operator implementations, compare precision with PyTorch implementations, and generate standardized precision reports.
Expert guidance for deep learning, transformers, diffusion models, and LLM development with PyTorch, Transformers, Diffusers, and Gradio.
Distributed training orchestration across clusters. Scales PyTorch/TensorFlow/HuggingFace from laptop to 1000s of nodes. Built-in hyperparameter tuning with Ray Tune, fault tolerance, elastic scaling. Use when training massive models across multiple machines or running distributed hyperparameter sweeps.
Identify and eliminate host-device synchronizations in PyTorch code. Detects sync points (.item(), .cpu(), boolean indexing, torch.tensor on CUDA), classifies false vs true dependencies, provides sync-free alternatives. Triggers: sync-free, synchronization, .item(), .cpu(), host-device sync, eliminate syncs, CPU stall, non_blocking, set_sync_debug_mode, cudaStreamSynchronize, cudaEventSynchronize, remove syncs, async GPU.
Provides guidance for interpreting and manipulating neural network internals using nnsight with optional NDIF remote execution. Use when needing to run interpretability experiments on massive models (70B+) without local GPU resources, or when working with any PyTorch architecture.
Guidance for implementing tensor parallelism in PyTorch, including ColumnParallelLinear and RowParallelLinear layers. This skill should be used when implementing distributed tensor parallel operations, sharding linear layers across multiple GPUs, or simulating collective operations like all-gather and all-reduce for parallel computation.
High-level PyTorch framework with Trainer class, automatic distributed training (DDP/FSDP/DeepSpeed), callbacks system, and minimal boilerplate. Scales from laptop to supercomputer with same code. Use when you want clean training loops with built-in best practices.
Generate a source-backed starting `trtllm-serve --config` YAML for basic aggregate single-node PyTorch serving, aligned with checked-in TensorRT-LLM configs and deployment docs. Preserves explicit latency / balanced / throughput objectives. Excludes disaggregated, multi-node, and non-MTP speculative configs.
LeetCode-style PyTorch interview practice environment with auto-grading for implementing softmax, attention, GPT-2 and more from scratch.
AI and ML expert including PyTorch, LangChain, LLM integration, and scientific computing
Debug PyTorch 2 compiler stack failures including Dynamo graph breaks, Inductor codegen errors, AOTAutograd crashes, and accuracy mismatches. Use when encountering torch.compile errors, BackendCompilerFailed exceptions, recompilation issues, Triton kernel failures, FX graph problems, or when the user mentions debugging PT2, Dynamo, Inductor, or compiled model issues.
Apply CUDA Graphs to PyTorch workloads — API selection (torch.compile, PyTorch make_graphed_callables, TE make_graphed_callables, MCore CudaGraphManager, FullCudaGraphWrapper, manual torch.cuda.graph), code compatibility, capture workflows, dynamic pattern handling, and troubleshooting. Triggers: CUDA graph, torch.cuda.graph, make_graphed_callables, reduce-overhead, graph capture, graph replay, kernel launch overhead, CudaGraphManager, FullCudaGraphWrapper, full-iteration graph, stream capture.