COAP is a memory-efficient training method that reduces computational overhead without sacrificing performance.
-
Updated
Mar 25, 2025 - Python
#
memory-efficient-training
Here are 8 public repositories matching this topic...
Memory-efficient PyTorch optimizers with Adaptive Log-Space quantization, fused CUDA kernels, and Adafactor, CAME, APOLLO support.
machine-learning deep-learning cuda pytorch quantization memory-efficient adam-optimizer fine-tuning low-rank optimizers low-bit llm-training 8bit-optimizer adafactor memory-efficient-training optimizer-research apollo-optimizer came-optimizer optimizer-quantization
-
Updated
Oct 16, 2026 - Python
Finite-state counter synapses + reversible activations for memory-efficient training. 6-bit code holds the entire per-parameter training state — no FP master weight, no Adam moments. 1.21B params trains on a single T4 in 2.25 GiB peak.
research deep-learning optimizer pytorch reversible quantization ternary low-resource bitnet memory-efficient-training
-
Updated
Nov 7, 2026 - Python
APOLLO optimizer experiments for LLM training: optimizer-state memory and convergence on TinyStories.
-
Updated
Oct 11, 2026 - Python
AAMLA: An Autonomous Agentic Framework for Memory-Aware LLM-Aided Hardware Generation
hardware memory gpu eda computer-aided-design gpu-memory-profiler large-language-models llm parameter-efficient-fine-tuning agentic-workflow agentic-ai memory-efficient-training estimation-tool
-
Updated
Mar 22, 2026 - Python
Patch-based training and inference for CountGD open-world object counting on an 11 GB GPU (B.Tech, IIIT-Delhi)
-
Updated
Nov 5, 2026 - Python
PyTorch implementation of Zero-Gated Cross-Attention (IGCA), a zero-initialized bridge that lets a frozen LLM take in images without changing its outputs at the start of training. Runs on a 4-bit Falcon-40B across multiple GPUs, verifies an exact 0.00e+00 initialization identity, and cuts held-out loss by 48.9%.
pytorch quantization multimodal efficient-training gradient-checkpointing ml-systems large-language-models cross-attention bitsandbytes parameter-efficient-fine-tuning model-sharding 4-bit-quantization memory-efficient-training frozen-llm
-
Updated
Nov 5, 2026 - Python
RetainCode: One-Bit Exact Gated-MLP Activation Compression from Retained Products DOI: https://doi.org/10.5281/zenodo.21698623
floating-point preprint swiglu memory-efficient-training activation-compression gated-mlp one-bit-encoding exact-gradients
-
Updated
Aug 23, 2026 - TeX
Add this topic to your repo
To associate your repository with the memory-efficient-training topic, visit your repo's landing page and select "manage topics."