Menu
Skills dans ce dépôt
SkillsMP a collecté 1 228 skills depuis ADu2021/skillXiv. Ouvrez un skill pour examiner sa source et ses détails.
ADu2021/skillXivAffichage de 40 skills collectés sur 1 228.
Enhances RLVR through cooperative-adversarial flywheel, intent verification, and trap instructions. Detects reward hacking and improves training efficiency, achieving 87.43% on IFEval.
Langue du texte source : anglais
Evaluate reasoning model capabilities by analyzing three complexity-dependent behavioral regimes and identifying fundamental limitations in symbolic manipulation rather than computational budgets.
Langue du texte source : anglais
Upscale any degraded image to 4K using an agentic framework that analyzes image quality, selects appropriate restoration tools, and iteratively improves results through reasoning and reflection.
Langue du texte source : anglais
Imagine-then-Plan: Agent Learning from Adaptive Lookahead with World Models. From arXiv:2601.08955
Langue du texte source : anglais
Accelerate sparse attention by reusing token selection indices across layers. Partition layers into full indexer (F) and shared (S) types using greedy search or multi-layer distillation to eliminate 75% of indexer computation.
Langue du texte source : anglais
Introduce IMDD-1M dataset with 1M aligned image-text pairs spanning 60+ material categories and 400+ defect types. Train diffusion-based vision-language models requiring 5% task-specific data vs. dedicated expert models for manufacturing quality control.
Langue du texte source : anglais
Implement techniques from Inference-Time Scaling of Verification: Self-Evolving Deep Research Agents via Test-Time Rubric-Guided Verification. While the majority of existing efforts focus on enhancing policy capabilities via post-training, we propose an…
Langue du texte source : anglais
Combines SFT and DPO with robust data selection pipeline using multidimensional quality metrics. Achieves DeepSeek-R1 performance with 12% training data, enabling efficient reasoning model alignment.
Langue du texte source : anglais
Research contribution advancing agent and reasoning capabilities through novel approaches to model development, training, and evaluation.
Langue du texte source : anglais
Merge sliding window and linear attention (Gated DeltaNet) for unlimited VLM inputs with 3.6× speedup. InfiniteVL handles video understanding at 24 FPS with constant memory—ideal when context length must scale without quadratic overhead.
Langue du texte source : anglais
Implement dense-sparse switchable attention enabling LLMs to scale from short to long sequences with 4× speedup and 98-99.7% performance retention, requiring no extra parameters by reusing pretrained attention weights through trainable sparse pattern…
Langue du texte source : anglais
Optimize multi-turn agent policies by measuring turn-level information gain via counterfactual reasoning. Provide dense reward signals identifying which clarifying questions and observations improve the agent's decision distribution, then adaptively blend…
Langue du texte source : anglais
Enable more precise LLM error correction by having users directly edit the model's previous response, conditioning the next response on this corrected version. This approach reduces token overhead by 79% compared to traditional separate-feedback methods while…
Langue du texte source : anglais
Enable VLMs to perform generalized visual search—locating relational, fuzzy, and conceptual regions from free-form language descriptions. Introduces O3-Bench benchmark with high-density composite charts/maps, uses RL-trained vSearcher for spatial…
Langue du texte source : anglais
Convert pre-trained latent video diffusion into real-time autoregressive generators using adversarial post-training, achieving 24fps streaming on single H100.
Langue du texte source : anglais
Enhance multimodal models through cascade RL for reasoning improvement and visual resolution routing for inference efficiency, achieving 16% reasoning gains and 4.05x speedup.
Langue du texte source : anglais
Map text embeddings into visual representation space for multimodal fusion, eliminating expensive image-text alignment pre-training while improving reasoning-heavy tasks by up to 27.2%.
Langue du texte source : anglais
Connect LLM agents to IoT sensors and microcontrollers through MCP standardization. Use to build monitoring systems and smart home automation where LLMs reason over real-world sensor data.
Langue du texte source : anglais
Enable efficient parallel decoding while preserving causal inference properties through progressive distillation with cyclic noise scheduling. Reduce training complexity via block-wise sparse attention, implement rejection recycling and multi-block decoding.…
Langue du texte source : anglais
Enable scalable MoE inference by disaggregating attention and expert layers onto independent GPU sub-clusters. Use adaptive two-phase communication, activation load-balanced scheduling, and activation-aware expert management. Achieve 3.9× higher per-GPU…
Langue du texte source : anglais
Implement techniques from Jet-RL: Enabling On-Policy FP8 Reinforcement Learning with Unified Training and Rollout Precision Flow. Reinforcement learning (RL) is essential for enhancing the complex reasoning capabilities of large language models (LLMs)
Langue du texte source : anglais
Accelerate diffusion transformers by processing only sparse anchor tokens in early stages, then expanding to full spatial resolution using learnable extrapolation. Combines SAG-ODE velocity field lifting with importance-guided token activation for lossless…
Langue du texte source : anglais
Demonstrate that simple single-stage RL with fixed hyperparameters matches complex multi-stage approaches for training small LLMs on mathematical reasoning. Use basic setup: GRPO algorithm, rule-based verification, 16K token context, standard training data…
Langue du texte source : anglais
Generate optimized GPU kernels by treating LLMs as planning engines that co-evolve with a world model. Decouples high-level algorithmic planning from low-level implementation, enabling structured search through optimization strategies. LLM world model…
Langue du texte source : anglais
Pixel-based reinforcement learning agents often fail under purely visual distribution shift even when latent dynamics and rewards are unchanged, but existing benchmarks entangle multiple sources of shift and hinder systematic analysis. We introduce KAGE-Env,…
Langue du texte source : anglais
Automate compute kernel optimization via agentic AI and retrieval-augmented prompting. Unified context-aware transformation function, hardware-specific constraint KB, self-improving state machine exploring kernel variants—achieving 1.25-17× speedups on…
Langue du texte source : anglais
Reveals that memory bandwidth—not computation—dominates test-time costs; proposes eFLOPs metric incorporating both computation and memory, showing 14B+ minimum threshold for reasoning value.
Langue du texte source : anglais
Generative approach for creating complex dynamic scenes and content, supporting agent capabilities in understanding and reasoning about multi-agent environments.
Langue du texte source : anglais
Unify video generation, editing, and reasoning through Multimodal Visual Language (MVL) paradigm processing text instructions, reference images, and video contexts in shared embedding space. Employ prompt enhancer, omni-generator with diffusion transformer,…
Langue du texte source : anglais
Improve RL training for LLMs by dynamically allocating exploration budget (rollout count) to tasks based on their difficulty and current learning status. Solves the knapsack problem of maximizing gradient signal within fixed compute budget, increasing…
Langue du texte source : anglais
Enable real-time portrait animation via causal AR video generation with temporal coherence. Sliding window with global reference frame caching, temporal knot module overlapping adjacent chunks, global context running ahead—preventing error accumulation while…
Langue du texte source : anglais
Build knowledge agents that generalize across grounded reasoning tasks by combining agentic data synthesis via vector search with off-policy reinforcement learning. Use multi-task training on synthetic question-answer pairs to achieve robust cross-domain…
Langue du texte source : anglais
Enables 8x KV cache compression with minimal training overhead to improve reasoning accuracy by allowing more token generation within computational budgets.
Langue du texte source : anglais
Guide frozen language models toward multi-step reasoning by modifying cached key-value representations after the prefilling stage. Extract steering vectors from contrastive prompt pairs and apply them to KV cache with scalar coefficients. Improves reasoning…
Langue du texte source : anglais
Extract high-quality embeddings from frozen decoder-only LLMs by re-routing internal key-value states without training—outperforming training-free baselines by 10% on MTEB while maintaining robustness across sequences up to 4,096 tokens.
Langue du texte source : anglais
Build an 8B multimodal model specializing in short-form video understanding. Combines four-stage pre-training with instruction-tuning and reinforcement learning to enable advanced reasoning about dynamic video content while maintaining general vision-language…
Langue du texte source : anglais
Enable LLM agents to actively explore and adapt policies through meta-RL instead of converging to fixed behaviors. Uses cross-episode training with trajectory-level discounting and in-context policy adaptation via textual reflections—achieving 11-19%…
Langue du texte source : anglais
Align vision-language-action models with natural language instructions using Bayesian decomposition with latent action queries, improving robotic manipulation generalization. Use when building agents that follow language commands while observing visual scenes…
Langue du texte source : anglais
Advanced reasoning approach for optimizing inference efficiency through meta-cognitive planning, enabling agents to make better decisions with reduced computational overhead.
Langue du texte source : anglais
Compute reasoning rewards from the model's own next-token probability distribution at solution end. Integrates verification-based feedback into single model without separate evaluator, enabling efficient RL training with minimal overhead.
Langue du texte source : anglais