All authors
Claude Skills by wenmin-wu
github.com/wenmin-wu535 skills0 installs688 views
- Slice Location Weighted PriorEncode the empirical per-position event rate as a soft prior over slice index by binning normalized slice location and looking up a precomputed weight vector — adds class-aware spatial context to a per-slice classifier without changing the modelVotes: 0GitHub stars: 61
- Slice Padding Augmented DuplicatesPad 3D volumes with fewer slices than required by duplicating existing slices with slight brightness variation via convertScaleAbsVotes: 0GitHub stars: 61
- Smm Spatial Observation EncodingEncode game state as a Super Mini Map (SMM) with separate binary channels for players, ball, and ownership, bit-packed for efficient transfer in RL trainingVotes: 0GitHub stars: 61
- Snapshot Cosine EnsembleCyclic cosine annealing LR that produces M diverse snapshots from a single training run for free ensemblingVotes: 0GitHub stars: 61
- Soft Threshold Count AggregationAggregate patch-level count-regression predictions into image-level totals by multiplying each prediction by a boolean mask (x * (x > tau)) instead of rounding, retaining fractional evidence above threshold while killing background noiseVotes: 0GitHub stars: 61
- Sparse Event Temporal Window ExpansionTurn sparse single-frame event labels (impacts, collisions, goals) into a usable detector training set by stamping the positive label onto a ±k-frame window around each event, then dropping any frame that contains no positives — gives the detector enough positive samples without changing the original annotationVotes: 0GitHub stars: 61
- Spatial Rect Train Val SplitSplit large-image segmentation data into train/val by spatial rectangle regions with border buffer exclusion to prevent patch leakageVotes: 0GitHub stars: 61
- Spectral Index SegmentationCompute normalized spectral band ratios (NDWI, CCCI, NDVI) from multispectral imagery and threshold for binary segmentation of water, vegetation, or other targetsVotes: 0GitHub stars: 61
- Stochastic Weight AveragingKeras callback implementing Stochastic Weight Averaging (SWA) — running mean of model weights over final training epochsVotes: 0GitHub stars: 61
- Stratified Fold By Sequence LengthStratifies cross-validation folds by output sequence length to ensure balanced length distributions across train/val splits.Votes: 0GitHub stars: 61
- Stroke Normalize Simplify PipelineNormalize raw stroke coordinates to 0-255 range, resample at uniform arc-length spacing, then apply Ramer-Douglas-Peucker simplificationVotes: 0GitHub stars: 61
- Stroke Temporal Color RenderingRender stroke sequences to grayscale images with temporal intensity encoding where earlier strokes are brighter and later strokes fade to encode drawing orderVotes: 0GitHub stars: 61
- Structured Gt Serialization TokensSerialize structured ground truth (chart data, table rows, form fields) as special-token-delimited sequences for generative vision-language model trainingVotes: 0GitHub stars: 61
- Temporal Frame Jitter AugmentationAdd random temporal offset to the center frame during training to augment temporal diversity in video-based modelsVotes: 0GitHub stars: 61
- Tensor Core Aligned PaddingPad batch sequence lengths to multiples of 8 for efficient tensor core utilization on GPUs, with -100 masking for label paddingVotes: 0GitHub stars: 61
- Tensorrt Fp16 CompilationCompiles a PyTorch model to a TensorRT FP16 engine via torch_tensorrt for 2-5x inference speedup, saved as reusable TorchScript.Votes: 0GitHub stars: 61
- Test Time AugmentationApplies geometric and color augmentations at inference time and averages predictions to reduce variance.Votes: 0GitHub stars: 61
- Tile Grid Mosaic AssemblyAssembles N extracted tiles into a sqrt(N) x sqrt(N) mosaic image for single-forward-pass CNN inference on whole slide images.Votes: 0GitHub stars: 61
- Tissue Content Tile SelectionSelects the top-N most informative tiles from a gigapixel whole slide image by ranking on pixel intensity sum, keeping tiles with the most tissue content.Votes: 0GitHub stars: 61
- Triple Stratified Shard FoldsPre-shard TFRecords into N files each balanced along 3 axes (patient, target, image-count), then KFold over file indices for leak-free triple-stratified foldsVotes: 0GitHub stars: 61
- Tumor Volume Ratio FeaturesExtract volumetric features from 3D segmentation masks including scan/tumor pixel ratios, tumor percentage, and tumor centroid coordinatesVotes: 0GitHub stars: 61
- Tversky LossTversky loss with independent alpha/beta constants to separately penalize false positives and false negatives in imbalanced segmentation.Votes: 0GitHub stars: 61
- Two Stage Bce Then LovaszTrain segmentation with BCE loss first for stable convergence, then fine-tune with Lovasz-hinge on raw logits for IoU-optimal predictionsVotes: 0GitHub stars: 61
- Two Stage Classify Then DetectChain a study-level classifier with an image-level detector, merging class probabilities and bounding boxes into a unified predictionVotes: 0GitHub stars: 61
- Two Stage Slice Localization Then Organ CropUse a lightweight slice-classifier to find the Z-range containing an organ in a CT volume, then crop and trilinear-resample that sub-volume into a fixed shape for a heavier 3D classifier — replaces "use the whole volume" with "use only the relevant slab" at a fraction of the FLOPsVotes: 0GitHub stars: 61
- Video Frame Sampling PipelineEfficiently sample N evenly-spaced frames from a video using OpenCV grab/retrieve pattern with optional resize for batch face detection or classificationVotes: 0GitHub stars: 61
- Vqa Yes No Logit ScoringExtracts calibrated yes/no probabilities from a VQA model by masking all logits except yes/no token variants and renormalizing via softmax.Votes: 0GitHub stars: 61
- Warmup Cosine Lr ScheduleCombines gradual learning rate warmup with cosine annealing decay for stable fine-tuning of pretrained models.Votes: 0GitHub stars: 61
- Weak Supervision Label PropagationPropagates image-level multi-labels to individual instances as weak supervision for instance-level training.Votes: 0GitHub stars: 61
- Weighted Bce LossBCE loss with per-class asymmetric positive/negative weights to match competition metrics or handle class imbalance in multilabel classification.Votes: 0GitHub stars: 61
- Weighted Embedding BlendingEnsemble predictions from heterogeneous vision-language models by blending their output embeddings with fixed scalar weights in embedding spaceVotes: 0GitHub stars: 61
- Yolo Bbox Format ConversionConvert bounding boxes between YOLO (normalized center), VOC (absolute corners), and COCO (absolute xywh) formats with image dimension scalingVotes: 0GitHub stars: 61
- 4bit Nf4 Double QuantizationLoad large LLMs with 4-bit NF4 quantization and optional double quantization via BitsAndBytes to reduce GPU memory by 4x while preserving inference qualityVotes: 0GitHub stars: 61
- Actor Critic Game AgentShared-backbone neural network with actor (policy) and critic (value) heads for grid-based game agent RL trainingVotes: 0GitHub stars: 61
- Adaptive Time Budget InferenceDynamically reduce max_tokens and batch size as wall-clock time approaches a cutoff to ensure all inputs get processedVotes: 0GitHub stars: 61
- Agreement Confidence EnsembleEnsemble multi-model LLM predictions using weighted combination of average probability, cross-model agreement ratio, and max confidenceVotes: 0GitHub stars: 61
- Answer Accumulation FilterParse sequential yes/no answers to build inclusion/exclusion sets, then apply compound boolean filters to narrow a candidate listVotes: 0GitHub stars: 61
- Assistant Prefill Response SteeringSteers LLM output format by injecting a partial assistant response into the chat template before generation, forcing structured output without fine-tuning.Votes: 0GitHub stars: 61
- Batched Perplexity ScoringBatch-compute perplexity for multiple texts using a causal LM with proper padding, shifted labels, and pad-token masking for efficient GPU utilizationVotes: 0GitHub stars: 61
- Binary Answer ClampingForce-clamp free-form LLM output to binary yes/no with keyword matching and a fallback default for constrained environmentsVotes: 0GitHub stars: 61
- Binary Search Entity NarrowingHierarchical binary search over entity space by asking category, region, then first-letter questions to narrow candidates before guessingVotes: 0GitHub stars: 61
- Boxed Answer ExtractionExtract final numeric answers from LaTeX \boxed{} notation in LLM math reasoning output, scanning matches in reverse for robustnessVotes: 0GitHub stars: 61
- Cargo Threshold State MachinePer-agent COLLECT/DEPOSIT state machine driven by cargo thresholds with greedy neighbor selection for resource collection gamesVotes: 0GitHub stars: 61
- Code Execution Weighted VoteExecute Python code blocks from LLM math responses in a sandbox, then double-weight code-derived answers in majority votingVotes: 0GitHub stars: 61
- Completion Only Lm TrainingFine-tune an instruct LLM on a single text field containing prompt+answer using TRL's DataCollatorForCompletionOnlyLM, which masks the loss to only the answer tokens by detecting a response template string at collation timeVotes: 0GitHub stars: 61
- Confidence Threshold FallbackUses primary model predictions only when confidence exceeds a threshold, falling back to a backup ensemble otherwise.Votes: 0GitHub stars: 61
- Counting Trap Adversarial EssayEmbed an enumerated list plus a counting/identity instruction to derail LLM judges into meta-answering instead of scoringVotes: 0GitHub stars: 61
- Dual Role Agent DispatchSingle agent entry point that dispatches between multiple roles (ask/answer/guess) based on turn type, combining heuristic and LLM-based strategiesVotes: 0GitHub stars: 61
- Exhaustive Permutation Early StoppingEnumerate all factorial permutations in batches with LLM scoring, tracking the running best and early-stopping when score crosses a known optimality thresholdVotes: 0GitHub stars: 61
- Few Shot Conversation InjectionInject fake (user, assistant) turn pairs before the real prompt to steer the LLM into a specific output formatVotes: 0GitHub stars: 61