📖 This is a repository for organizing papers, codes, and other resources related to Latent Reasoning.
-
Updated
Nov 5, 2025
#
diffusion-llm
Here are 24 public repositories matching this topic...
Speed Always Wins: A Survey on Efficient Architectures for Large Language Models
moe mamba linear-models state-space-model mixture-of-experts efficient-architectures linear-attention sparse-attention linear-rnn diffusion-llm
-
Updated
Nov 11, 2025
paper list, tutorial, and nano code snippet for Diffusion Large Language Models.
diffusion diffusion-models llm diffusion-lm diffusion-language-models masked-diffusion-lm diffusion-llm large-diffusion-language-models diffusion-large-language-model auto-regressive-diffusion-model block-diffusion large-language-diffusion-model dllm
-
Updated
Jan 19, 2026 - Jupyter Notebook
Experimental playground for benchmarking language model (LM) architectures, layers, and tricks on smaller datasets. Designed for flexible experimentation and exploration.
machine-learning deep-learning transformers pytorch gpt attention-mechanisms gpt-2 position-embedding large-language-models llm llm-training hymba diffusion-llm dllm
-
Updated
Oct 2, 2026 - Python
[ICLR 2026] ParallelBench: Understanding the Tradeoffs of Parallel Decoding in Diffusion LLMs
-
Updated
Mar 27, 2026 - Python
Educational Implementation of "Edit Flows: Flow Matching with Edit Operations" by Havasi et al.
-
Updated
Oct 17, 2025 - Jupyter Notebook
[MIR🎉] Quantization Meets dLLMs: A Systematic Study of Post-training Quantization for Diffusion LLMs
-
Updated
Apr 24, 2026 - Python
[ACL '26] UNCODE (a.k.a. PC-Sampler): training-free decoding calibration for masked diffusion LLMs. Code for "Empirical Analysis of Decoding Biases in Masked Diffusion Models"
nlp decoding text-generation sampling reasoning uncode non-autoregressive llm training-free discrete-diffusion llada diffusion-language-models diffusion-llm dllm acl2026 masked-diffusion-models pc-sampler
-
Updated
Oct 24, 2026 - Python
Anchored Diffusion Language Model (NeurIPS 2025)
planning generative-modeling large-language-models auto-regressive-models discrete-diffusion chain-of-thought-reasoning diffusion-llm
-
Updated
Oct 13, 2025 - Python
Decisions, not essays: a choice, a yes or no, or a score with a calibrated probability. Open weights, Jev-compatible API, adaptive thinking.
calibration classification decision-model ai-agents mixture-of-experts guardrails vllm llm-routing diffusion-llm open-weights jev system-one
-
Updated
Oct 26, 2026 - Python
A Curated Collection of Frontier Language Model Architectures
-
Updated
Jan 30, 2026
[ACL 2026 Main] Code for paper "d-TreeRPO: Towards More Reliable Policy Optimization for Diffusion Language Models"
-
Updated
Jan 6, 2026 - Python
DeepSeek Harness (dsh) plugin for Inception Mercury: renders the diffusion denoising stream live inside the answer bubble
-
Updated
Oct 26, 2026 - TypeScript
[ICLR2026] Official Code for Loopholing Discrete Diffusion: Deterministic Bypass of the Sampling Wall
text-generation discrete-diffusion diffusion-llm diffusion-language-model iclr2026 self-conditioning uniform-state-diffusion
-
Updated
Aug 10, 2026 - Python
A curated list of diffusion language models (dLLMs): papers, models, and code behind diffusion-based, non-autoregressive text generation, focused on 2025–2026.
awesome deep-learning text-generation awesome-list papers diffusion-models large-language-models llm diffusion-llm diffusion-language-model
-
Updated
Oct 29, 2026
FastMCP fleet MCP server for diffusion LMs (dLLM). DiffusionGemma on Goliath RTX 4090 — batch inference, HLE-shaped reasoning, ~200–400 tok/s. Doc phase; llama-diffusion-cli sidecar next. Complements local-llm-mcp.
python windows mcp inference gemma diffusion batch-inference llm llama-cpp local-llm discrete-diffusion gguf agentic-ai model-context-protocol mcp-server fastmcp diffusion-llm dllm open-weights diffusiongemma
-
Updated
Jun 17, 2026
Diffusion VLM papers indexed with one feature schema, plus decode-behaviour traces measured through a single instrumentation path
-
Updated
Aug 3, 2026 - Python
Experiment monorepo: speculative decoding (DFlash, DDTree on GDN hybrids, typical acceptance), prefix caching and long-context RoPE on one Jetson AGX Thor; 10 hypotheses with verdicts and evidence
mixture-of-experts long-context vllm llm-inference speculative-decoding diffusion-llm jetson-thor dflash
-
Updated
Oct 26, 2026 - Python
DiffusionGemma-26B-A4B NVFP4 on NVIDIA Jetson AGX Thor with vLLM — single-stream decode speed by turn length, prefix caching (79-83% faster warm turns), quality checks, requant headroom
-
Updated
Oct 26, 2026 - Shell
Pure-CUDA inference engine + OpenAI-compatible server for LLaDA2.2-mini (16B MoE diffusion LLM) on Jetson Orin Nano Super 8GB
-
Updated
Oct 26, 2026 - Python
Add this topic to your repo
To associate your repository with the diffusion-llm topic, visit your repo's landing page and select "manage topics."