Models

An MCP server that lets a coding agent generate synthetic training datasets, run LoRA fine-tuning, do policy-gradient reinforcement learning against a local judge model, merge adapters and evaluate results, all without leaving the chat.

• MCP servers
• Models

A training and serving stack for fine-tuning open code models such as StarCoder2 and DeepSeek-Coder into low-latency IDE completion models, aimed at low-resource languages and on-premises deployment that mainstream hosted copilots skip.

• Models

A from-scratch, hand-built decoder-only transformer training project, currently a 152-million-parameter model pretrained on about 15 billion tokens with RoPE, RMSNorm and SwiGLU, released as a Hugging Face-compatible checkpoint with a live demo.

• Models
• Tutorials

An independent structural validation report on two split GGUF releases of the Kimi K3 model, checking descriptor counts, identities, and shapes between quantization variants without downloading any tensor payload bytes.

• Models
• Papers

A post-training quantization method, presented at ICML 2026, that compresses large language model weights below one bit per parameter using mixture-of-rank binary bases while preserving accuracy.

• Models
• Papers

An offline-first interactive lab and long-form guide to Muse Glimmer, Meta's open-weight 30-billion-parameter agentic multimodal model, letting a reader explore agent loops, benchmarks, and memory sizing before downloading the model weights.

• Models
• Tutorials

A visual, node-based editor for composing PyTorch neural network architectures as typed, executable cards, with two-way sync to generated code, step-by-step tensor execution, and a constrained AI planner that proposes only valid graph changes.

• Models
• Design tools

An educational, from-scratch PyTorch implementation of Mamba, the selective state space model architecture, covering the selective scan mechanism, a Mamba block, and a complete language model built on it.

• Models
• Tutorials

A 27B coding and tool-use language model family for agentic repository work, released both as a full-precision PEFT adapter on a Qwen3.6 base and as a standalone, packed-weight GGUF edition for native local inference.

• Models
• Inference

A research framework for fine-tuning and evaluating ECG-language models, which combine electrocardiogram signals with text for tasks such as ECG question answering.

• Models

A companion toolkit that adds LoRA fine-tuning to CosyVoice v3 for single-speaker voice cloning on one consumer GPU, since the upstream code only supports full fine-tuning.

• Models
• Speech

A sparse fine-tuning method that trains a small fraction of a Mixture-of-Experts model's expert weights and router, making MoE fine-tuning possible on a 12GB consumer GPU.

• Models

A chat-driven LLMOps agent that turns a folder of documents into a fine-tuned, domain-specific language model, with human approval at each pipeline step.

• Models
• Data engineering

A lightweight, configuration-driven fine-tuning framework for Hugging Face Diffusers pipelines, with a simplified LoRA implementation and native distributed training.

• Models
• Generative media

A decoder architecture that folds self-attention into a recurrent feedback pathway, combining input-dependent attention routing with feedback-based recurrent aggregation to improve long-context handling beyond standard Transformers and state-space models such as Mamba.

• Models

Official code for K-Forcing, which distills an autoregressive language model into a push-forward model that decodes several tokens jointly in one forward pass instead of one token at a time, for faster batch inference.

• Models
• Inference

A Mixture-of-Experts language model implemented from scratch in native PyTorch and trained on two free Kaggle T4 GPUs, meant to teach modern LLM internals.

• Models
• Tutorials

A one-stop NLP platform for small teams that trains, evaluates, and deploys small encoder models such as BERT and BGE, and separately runs blind, human-and-AI-scored prompt evaluations for large model prompts, without fine-tuning large models itself.

• Models
• AI evals

An out-of-the-box fine-tuning framework, built on Hugging Face Transformers, PEFT, and DeepSpeed, for training large language models to perform binary classification on medical text such as disease-risk prediction.

• Models

An automated code-review command-line agent that reviews a change, applies fixes, and runs tests, recording every step as a structured trajectory used to train an open-weight code-review model.

• Models
• Code quality

A dataset-generation skill for AI coding assistants that turns topics, URLs, or raw files into supervised fine-tuning and preference-optimization training datasets, with verification, deduplication, and export pipelines.

• Agent skills
• Models

A complete neural network framework written in pure C, with no Python or CUDA toolkit dependency, that both runs and trains models, including loading and running GGUF files directly.

• Models
• Inference

A self-hosted inference stack for a cluster of NVIDIA DGX Spark machines that routes tasks across local and cloud models, then LoRA-trains the local models on its own routing history so the local share of work grows over time.

• Models
• Inference

A research training method and codebase where a student model learns from a teacher's demonstrations while a constraint limits how far it can drift, aimed at fine-tuning a model on a new task without forgetting old ones.

• Models
• Papers