Open-source ML releases as they ship. Indexed live by the Starship No. 27 crawl pipeline.
Showing the most recent 30 stories · updated continuously
This text describes a training checkpoint for a reinforcement learning model used in robotics, focusing on the current state inference process and its configuration.
huggingface · 2026-09-23The text discusses model downloads and inference providers but does not mention any technical concepts related to DSP, MATH, HARDWARE/SDR, or DAW/UI.
huggingface · 2026-09-23The model fine-tuned on T5-base generates a summary of Python function code by translating it into natural language.
huggingface · 2026-09-23A 7B-core decoder-only model with a 512K context window.
huggingface · 2026-09-23A tiny test-time-training state-space model with a transactional safety harness.
huggingface · 2026-09-23This text describes a repackaged version of the Anima ControlNet-LLLite model for use with ComfyUI, providing details on where to download and how to organize the files.
huggingface · 2026-09-23Sparse mixture-of-experts encoder-decoder model with RoPE and YaRN.
huggingface · 2026-09-23This repository is a NVIDIA Model Optimizer quantized derivative of dots-studio/dots.mocr, intended for GPU deployment with specific precision settings.
huggingface · 2026-09-23README.md file is empty, no technical concepts discussed.
huggingface · 2026-09-23Krea 2 is an open-source repackaging of model files for ComfyUI, providing various LoRA styles and workflows.
huggingface · 2026-09-23A new AI-based audio processing model with a large parameter count and high precision.
huggingface · 2026-09-23This text describes the repackaging of model files for a diffusion model called Boogu, which is used in the ComfyUI image generation tool.
huggingface · 2026-09-23No information provided.
huggingface · 2026-09-23No information provided.
huggingface · 2026-09-23An immersive close-up special effects model adapted to any style image-to-video tool, featuring a distortion-free, high-precision, and cinematic pupil depth effect.
huggingface · 2026-09-23PPO agent trained to play LunarLander-v2 using Stable-baselines3.
huggingface · 2026-09-23A robotic arm effect is used to move a camera left and right with precision and smooth motion, bringing the medium shot closer to a close-up of the character's face.
huggingface · 2026-09-23This text discusses the training process of a deep learning model using OpenPI, focusing on its architecture and configuration settings.
huggingface · 2026-09-23LoRA parameter interpolation between two LLM adapters for instruction fine-tuning.
huggingface · 2026-09-23No model card downloads available; tracking is not implemented.
huggingface · 2026-09-23LLM fine-tuning, LoRA parameter interpolation, adapter creation.
huggingface · 2026-09-23PPO agent trained to play CartPole-v1 with hyperparameters and evaluation metrics.
huggingface · 2026-09-23No model card downloads available; tracking is not implemented.
huggingface · 2026-09-23No information provided about technical concepts.
huggingface · 2026-09-23The text discusses the lack of downloads and inference providers for a specific AI model, indicating it is not actively used or supported.
huggingface · 2026-09-23This paper introduces a gated DeltaNet architecture that uses recurrent attention blocks to refine its output based on predictive uncertainty.
huggingface · 2026-09-23The text discusses the lack of downloads and inference providers for a specific model, indicating it is not currently deployed.
huggingface · 2026-09-23A 2-sentence summary of the key technical concepts.
huggingface · 2026-09-23Transformers model card for a language model pushed on the Hub.
huggingface · 2026-09-23LoRA parameter interpolation between two adapters for a fine-tuned Qwen model.
huggingface · 2026-09-23