Code for "LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding", ACL 2024
-
Updated
Jul 20, 2026 - Python
#
early-exit
Here are 26 public repositories matching this topic...
Dynamic per-token early exit for LLM inference. Skip layers tokens don't need
-
Updated
Sep 18, 2026 - Python
Artifact for "Apparate: Rethinking Early Exits to Tame Latency-Throughput Tensions in ML Serving" [SOSP '24]
-
Updated
Nov 21, 2024 - Python
[NeurIPS'25] FreqExit: Enabling Early-Exit Inference for Visual Autoregressive Models via Frequency-Aware Guidance
-
Updated
Dec 15, 2025 - Python
The blazing-fast, multimodal AI decision engine. Uses adaptive early-exit computing and attention rollout for sub-15ms explainable zero-shot classification
nlp machine-learning transformers pytorch decision-engine explainable-ai early-exit zero-shot-classification
-
Updated
Oct 4, 2026 - Python
Official PyTorch implementation of "LGViT: Dynamic Early Exiting for Accelerating Vision Transformer" (ACM MM 2023)
-
Updated
Nov 18, 2024 - Python
Code for paper "EdgeKE: An On-Demand Deep Learning IoT System for Cognitive Big Data on Industrial Edge Devices"
iot deep-neural-networks deep-learning inference resnet knowledge-distillation edge-computing edge-devices resnets inference-optimization early-exit jetson-nano branchynet edge-applications accuracy-requirements efficient-ai
-
Updated
Oct 8, 2026 - Python
Think Less, Act Early: Reinforced Latent Reasoning with Early Exit in Vision-Language-Action Models
reinforcement-learning robotics vla robotics-algorithms early-exit vision-language-action openvla latent-reasoning
-
Updated
Sep 2, 2026 - Python
This repository provides an implementation of Early Exit Neural Network (EENN) and provides a way to profile the processing time of the model.
-
Updated
Mar 19, 2025 - Python
Deployment-oriented early-exit toolkit for PyTorch vision models, with per-head calibration, fair benchmarks, and staged inference.
-
Updated
Jul 22, 2026 - Python
Relational Time Engine (RTE): runtime density regulation for compute-efficient transformer inference. Demonstrates up to 75% layer reduction with improved latency and throughput.
python benchmark machine-learning deep-learning transformer event-driven energy-efficiency ai-systems inference-optimization early-exit transformer-architecture green-ai ai-optimization runtime-systems efficient-ai energy-efficient-ai transformer-inference-overhead runtime-gating relational-time
-
Updated
Mar 12, 2026 - Python
VerDee: Vertical Deep Network on Mamba-2 — staged LoRA (shallow/mid/deep), early-exit routing, and domain experts. Pilot on 370M; 16GB GPU friendly. Research/experiment repo.
lora state-space-model peft early-exit huggingface qlora llm-finetuning mamba2 vertical-deep-network
-
Updated
May 25, 2026 - Python
Token-level early-exit Transformer with Gumbel-Softmax routing and a latency-aware joint loss — skips redundant layers for "easy" tokens while going deep on hard ones.
nlp research deep-learning transformers pytorch gumbel-softmax model-compression efficient-inference early-exit adaptive-computation mixture-of-depths pareto-frontier
-
Updated
Aug 19, 2026 - Python
Early-exit, full-model and static-baseline inference on RubikPi 3, with live-camera, HTP, latency and thermal evidence.
computer-vision gstreamer qualcomm onnx early-exit edge-ai model-optimization embedded-ml rubikpi qcs6490
-
Updated
Jul 31, 2026 - Python
Lev: Thinking Fast and Deep with Looped Decision Models. One checkpoint, multiple depths: System 1.5 decision models that think in loops, not tokens.
research pytorch early-exit huggingface adaptive-computation llm looped-transformers decision-models adaptive-depth jev jev-api jev-model jevbench system-1-5
-
Updated
Oct 14, 2026 - Python
Confidence-gated early-exit inference for Llama 3.1 8B, with distillation, calibration, segmented multi-GPU runtime, and evaluation.
-
Updated
Jul 10, 2026 - Python
Code and reproducibility materials for TopoAudit | IEEE CSCloud 2026 | Manuscript in preparation
-
Updated
Oct 13, 2026 - Python
Early exit inference framework for HuggingFace LLMs — skip unnecessary transformer layers when the model is already confident. Supports LLaMA, Mistral, Phi, Gemma, Qwen, Pythia and more.
deep-learning transformers pytorch language-models knowledge-distillation inference-optimization early-exit huggingface model-optimization llm-inference
-
Updated
Mar 29, 2026 - Python
LLM early exit whose error guarantee survives distribution shift: online conformal threshold fed by random self-audits (Qwen3-8B, tuned lens + free gate)
-
Updated
Oct 3, 2026 - Python
Operating-regime study of thermal-aware anytime (early-exit) inference on a Raspberry Pi 5: when early-exit selection controls temperature, and when it does not.
raspberry-pi reproducible-research inference energy-efficiency thermal early-exit edge-ai arm-cortex-a76 anytime-inference
-
Updated
Oct 13, 2026 - Python
Add this topic to your repo
To associate your repository with the early-exit topic, visit your repo's landing page and select "manage topics."