Python 排行榜

[ACL-IJCNLP 2021] Self-Supervised Multimodal Opinion Summarization

An alternative EQA paradigm and informative benchmark + models (BMVC 2019, ViGIL 2019 spotlight)

This repository contains the official code for our paper: Thinking Before Looking: Improving Multimodal LLM Reasoning via Mitigating Visual Hallucination

Awesome-HCI （Ubiquitous, LLM, MLLM, Agent, RAG, Embodied-AI, RLHF)

Automate the creation of soccer match highlights with the power of Generative AI and AWS. This solution leverages AWS Bedrock (Anthropic’s Claude 3 Sonnet model), AWS MediaConvert, Lambda, Step Functions and other AWS services to identify and compile exciting game moments without manual editing.

Search knowledge by what documents mean and how they look — not one or the other.

A Benchmark Dataset for Multimodal Scientific Fact Checking

EVA: Efficient Reinforcement Learning for End-to-End Video Agent

A stand-alone application with GUI for OpenAI's Whisper

Reproducible voice-AI benchmarking — TTS / STT latency and accuracy.

Self-hosted, OpenAI-compatible speech server: faster-whisper STT and Piper/Kokoro TTS. A local voice backend for Open WebUI, CLIs, and custom agents.

Taiwanese Hokkien (Taigi) speech-to-text transcriber - MediaTek Breeze-ASR-26 with faster-whisper, tuned for RTX 3050 4GB low-VRAM GPUs. Gradio UI, CLI, Docker, SRT/VTT/TXT/JSON.

DRIA (Deep Research and Intelligence Agent) is a fully local voice assistant that can hold real-time conversations while performing deep research in the background — powered by Firecrawl, Mistral AI, Perplexica, and LiveKit.

This is an OpenAI Whisper automatic speech recognition microservice

PyTorch toolkit for streaming speech recognition, speech translation and simultaneous translation based on fairseq.

Probing the limitations of multimodal language models for chemistry and materials research

Training-free KV cache compression via E8 lattice VQ. 2-bit KV that preserves retrieval (30/30 NIAH vs TurboQuant 0/30). Calibration-free, 9 architectures validated.

Unified KV-cache compression for LLM inference: 12 Python-native methods, guarded add-on composition and routing, analytical capacity simulation, Godzilla KVarN/TriAttention with SM86/SM89 qualification, CUDA weight sharing, and multi-GPU planning.

Kubernetes scanner that discovers LLMs running on vLLM and extracts their deployment and runtime facts.

Pretrained VGG-16 network as feature extractor for Object Recognition (Python, Keras, Scikit-Learn)