Results for “qwen3”

Trains language models with Direct Preference Optimization using preference pairs, covering DPOTrainer setup, dataset preparation, and beta tuning for stable preference learning without explicit reward models.

China LLM Gateway - Unified interface for Chinese LLMs including Qwen, DeepSeek, GLM, Baichuan. OpenAI compatible, one API Key for all models. Use when: the user needs model routing, provider setup, or Chinese LLM access guidance.

Expert en informatique quantique avancée (Qiskit, Cirq, algorithms, error correction, DZ research)

Routing guide for the built-in QwenWork Connector tools (mcp__qw-builtin__qw_query / mcp__qw-builtin__qw_action). Load ONLY right before calling them to view or manage QwenWork's OWN tasks/sessions or app configuration, or when a qw tool result explicitly asks. Before loading, always check whether another skill or tool can do the job — if so, use that instead. Unless the user explicitly asks, never use it to view skills, plugins, MCP servers, or third-party connectors. Never load in any non-essential scenario (content creation, PPT/docs, coding, research, web tasks), even when the topic is QwenWork itself. If in doubt, do not load.

Agent E3 - Mixed Methods Integration Specialist - Qual-Quant data integration and meta-inference. Covers joint display creation, integration strategies, and legitimation techniques.

Agent C3 - Mixed Methods Design Consultant Comprehensive mixed methods research design specialist covering sequential, concurrent, embedded, and multiphase designs with Morse notation. Core Capabilities: - Sequential Explanatory (QUAN → qual): Explain quantitative results - Sequential Exploratory (QUAL → quan): Develop instruments - Convergent Parallel (QUAN + QUAL): Comprehensive understanding - Embedded (QUAN(qual)): Secondary strand addresses different question - Multiphase: Long-term projects with iterative phases - Morse notation interpretation and recommendation

Quantize large language models to 4-bit using activation-aware weight quantization, achieving ~3x speedup with minimal accuracy loss for deployment on limited GPU memory.

Expert en IoT avancé (MQTT, edge, device management, digital twin, security, DZ infrastructure)

Generate and edit images using Alibaba Qwen-Image-2.0 models via the inference.sh CLI, with support for text-to-image, multi-image editing, and text rendering.

Generate images with Alibaba Qwen-Image-2.0-Pro via inference.sh CLI, with professional text rendering and fine-grained realism for posters, banners, and text-heavy designs.

Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.

Automates genome-wide association studies from genotype files to publication-ready results, running PLINK2 QC and REGENIE regression with Manhattan and QQ plots.

Train reinforcement learning agents using PPO, SAC, DQN, TD3, DDPG, and A2C algorithms with a scikit-learn-like API. Supports custom Gymnasium environments, vectorized environments, callbacks, and model persistence.

Train quantum circuits with automatic differentiation and build hybrid quantum-classical models using PennyLane, including VQE, QAOA, and integration with PyTorch, JAX, and TensorFlow.

Agent validation v4.3.0 — Make agents act effectively by disabling harmful actions, lowering gates, and injecting cross-run learning

Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.

Quantize large language models to 4-bit precision using activation-aware weight quantization, reducing memory footprint and speeding up inference with minimal accuracy loss.

Quantize large language models to 8/4/3/2/1-bit precision without calibration data, using multiple optimized backends and integrations with HuggingFace Transformers, vLLM, and PEFT/LoRA.

Build, test, run, and flash NightRun — a bare-metal, no_std Rust UEFI application that boots straight into a local LLM (Llama 3.2, Qwen3, or Granite 4.1) with no operating system underneath. Use when the user wants to build/flash a bootable NightRun USB or Raspberry Pi 5 SD image, convert a GGUF model into the `.nrm` container, run/debug the inference engine on the host or in QEMU/OVMF, or troubleshoot no_std kernel/tokenizer parity issues in the NightRun codebase. Triggers on: "nightrun", "boot into an LLM", "bare-metal LLM runtime", "UEFI LLM appliance", "nrconvert", "nrhost", "cargo xtask", "nrm model file", "flash a bootable LLM USB".

Expert en télécommunications avancées (5G, fiber, MVNO, IoT connectivity, network optimization, DZ context)

Implements the NOWAIT technique for efficient reasoning in R1-style LLMs. Use when optimizing inference of reasoning models (QwQ, DeepSeek-R1, Phi4-Reasoning, Qwen3, Kimi-VL, QvQ), reducing chain-of-thought token usage by 27-51% while preserving accuracy. Triggers on "optimize reasoning", "reduce thinking tokens", "efficient inference", "suppress reflection tokens", or when working with verbose CoT outputs.

Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.

Implements the NOWAIT technique for efficient reasoning in R1-style LLMs. Use when optimizing inference of reasoning models (QwQ, DeepSeek-R1, Phi4-Reasoning, Qwen3, Kimi-VL, QvQ), reducing chain-of-thought token usage by 27-51% while preserving accuracy. Triggers on "optimize reasoning", "reduce thinking tokens", "efficient inference", "suppress reflection tokens", or when working with verbose CoT outputs.

Kling 3.0 video generation on RunComfy. Kling 3.0 (also called Kling V3.0) is Kuaishou Technology's third-generation multi-shot video model with native synchronized audio and consistent character identity across shots. This skill covers all six Kling 3.0 endpoints, spanning three rendering tiers (Standard, Pro, 4K) and two modes (text-to-video, image-to-video). Calls runcomfy run kling/kling-3.0/<tier>/<mode> through the local RunComfy CLI. Triggers on "kling", "kling 3.0", "kling v3", "kling pro", "kling 4k", "kling text to video", "kling image to video", or any explicit ask to generate or animate with Kling 3.0.

Expert en Web3 avancé (DeFi, DAOs, smart contracts, tokenomics, Layer 2, DZ regulations)