This repository provides a comprehensive, production-grade training framework for fine-tuning the Qwen-VL family of vision-language models (Qwen2-VL, Qwen2.5-VL, Qwen3-VL, Qwen3.5) using HuggingFace Transformers, Liger-Kernel, and DeepSpeed. It supports Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and classification fine-tuning. The architecture is modular: dataset loaders handle mixed-modality (image/video/text) data with reasoning field support, custom trainer subclasses optimize memory and enable generation-based evaluation, and model wrappers provide sequence classification heads with configurable MLP bridges. Key engineering achievements include monkey-patched vision transformer forward passes for reduced memory usage, robust handling of DeepSpeed ZeRO-3 with dummy vision inputs for text-only batches, and a workaround for PEFT/DeepSpeed compatibility. The codebase is actively maintained, dependencies are pinned, and the project demonstrates strong craftsmanship with several reusable components.
TinyLLaVA Factory is a modular Python framework for building, training, evaluating, and serving small-scale Large Multimodal Models (LMMs). It provides a pluggable architecture with separate components for vision towers (CLIP, SigLIP, DINOv2), language models (Phi, TinyLlama, Gemma, Qwen2, etc.), and connectors (MLP, QFormer, Resampler), all accessible via factory registries. The training pipeline supports two-stage pretraining (alignment) and finetuning with full, LoRA, or QLoRA recipes using DeepSpeed. It includes a comprehensive evaluation suite for standard benchmarks (VQA, GQA, SQA, POPE, MMMU, etc.) and a Gradio-based web demo. The codebase is well-organized and follows clean architectural patterns, though many modules are tightly coupled to the internal framework. It is actively maintained, with recent additions like a visualization tool and support for new models.
This repository is a curated survey of research papers on World Action Models (WAM), maintained as a living list in a single README.md file. It categorizes 109 papers into three families: Render-and-Decode, Latent-Only, and Video-Generation-Free, with each entry including publication date, title, backbone, substrate, and a one-line summary. The repository serves as a reference resource for researchers in robotics and embodied AI, providing a structured overview of recent advances in WAM methodologies. While it contains no executable code or data processing pipelines, it is actively updated via community contributions through pull requests.
The openai/CLIP repository provides a Python library for loading and using CLIP (Contrastive Language-Image Pre-Training) models. It offers a clean API for model loading, image preprocessing, text tokenization, and computing similarity between images and text. The core implementation includes two vision encoder architectures (ModifiedResNet and VisionTransformer) and a Transformer-based text encoder, all built with PyTorch. The library is designed for research and inference, enabling zero-shot image classification, image-text retrieval, and feature extraction. While primarily a research artifact, the code is well-structured and demonstrates strong engineering practices, with reusable components such as the BPE tokenizer and model architecture definitions. The repository is stable but not actively developed, with dependencies on older PyTorch versions.
The repository provides the official PyTorch implementation of DoRA (Weight-Decomposed Low-Rank Adaptation), a parameter-efficient fine-tuning method. It integrates with Hugging Face PEFT and includes training/evaluation scripts for vision-language and language tasks (e.g., VQA, ScienceQA, commonsense reasoning). The architecture is modular, with core components like DoRA tuners, adapter controllers, and low-rank layers. Entry points include training scripts (finetune.py, train_dora.py), evaluation scripts (evaluate.py, run_llava.py), and serving tools (gradio_web_server.py). External dependencies include PyTorch, Transformers, PEFT, bitsandbytes, and dataset-specific libraries.
DeepSeek-OCR is an inference repository for a multimodal OCR model that combines vision encoders (SAM and CLIP) with a large language model (DeepSeek-V2/V3) to perform optical character recognition with layout awareness, supporting native and dynamic resolution processing. The system provides two inference paths: a Hugging Face Transformers integration for simple usage and a custom vLLM integration for high-throughput batched inference, including PDF processing and evaluation benchmarks. The architecture is a monolithic model definition with separate modules for vision encoding (deepencoder), image preprocessing (process), and vLLM glue code, all tightly coupled to the specific model architecture. The codebase is functional and demonstrates competent engineering for research-grade inference, but lacks production hardening, documentation, and consistent coding standards.
LLaVA-NeXT is a large multimodal model (LMM) research codebase that extends the LLaVA family to support single-image, multi-image, video, and interleaved tasks. It provides training pipelines for supervised fine-tuning (SFT), direct preference optimization (DPO), and reinforcement learning (PPO, GRPO) with custom vision encoders (CLIP, SigLIP, EVA-CLIP, MLCD) and modular projector/resampler components. The repository also includes a full RL training framework (EasyR1/verl) for large-scale GRPO training using Ray, FSDP, and vLLM, as well as evaluation scripts for video and interleaved benchmarks. While primarily a research codebase with extensive training configurations and experimental scripts, it also contains serving infrastructure (Gradio, SGLang, controller-worker) for deployment. The code is well-structured with reusable vision encoder wrappers, data protocols, and RL algorithm implementations, but suffers from hardcoded internal paths, multiple overlapping RL implementations, and legacy training pipelines that increase maintenance burden.
Show-o and Show-o2 are research repositories for unified multimodal understanding and generation. They implement transformer-based models (1.3B, 1.5B, 7B parameters) that can perform text-to-image generation, image/video captioning, visual question answering, inpainting/extrapolation, and mixed-modality generation (interleaved text and images). The codebase is organized around two major model versions: Show-o uses discrete tokens with MAGVITv2 VAE and Phi-1.5 LLM, while Show-o2 moves to continuous latent space with WanVAE (3D causal VAE) and Qwen2.5 LLM. Training follows a multi-stage pipeline (pretraining, instruction tuning) with novel techniques like flow matching (via a modular `transport` package), dual-path vision encoders (SigLIP + patch embedding), and omni-attention masks that combine causal and bidirectional attention for multimodal sequences. The repository includes distributed training configs (DeepSpeed, multi-node), evaluation scripts for GenEval and DPG-Bench, and comprehensive YAML configuration files. This is primarily a data_research codebase: it produces research artifacts (models, papers at ICLR 2025 and NeurIPS 2025) rather than a production application. The code exhibits good modularity for a research project, with reusable components like the transport module, omni-attention masks, mixed dataloader, and attention masking utilities. However, it suffers from typical research-code issues: hardcoded internal infrastructure paths, reliance on non-public data pipelines (e.g., internal packages for RefinedWeb), and some code duplication between Show-o and Show-o2. Reproducibility is moderate—dependencies are pinned and configs are detailed, but many datasets are large-scale web sources that require manual setup, and the training scripts assume specific hardware (multi-node A100 clusters). The project is actively maintained (commits in early 2026) with a roadmap for future work, including mixed-modal generation and video support.
This repository serves as a companion to the survey paper 'Towards Open Vocabulary Learning: A Survey' published in T-PAMI 2024. It contains two markdown files: README.md and dataset.md. The README provides a curated, tabular listing of papers covering open-vocabulary object detection, segmentation, video understanding, and 3D scene understanding, along with related domains. The dataset.md lists standard benchmarks and dataset splits used across these tasks. The repository does not contain any executable code, data processing pipelines, or model implementations. It is purely a literature survey resource. While useful as a reference for researchers, it has no reusable software components and no data lineage to trace.