ternary-weights

A 1.6 MB model using ternary 3:4 weights runs on WebGPU in the browser, playing Connect Four as well as a 7.8 MB int8 model, demonstrating the effectiveness of ternary formats for lightweight AI deployment.

A controlled re-examination of ternary language models at 60K parameters reveals that baseline shape significantly impacts performance comparisons, challenging previous claims about the routed ternary model's advantage.

A Twitter post discusses running the uncensored Ternary-Bonsai-27B AI model locally on a MacBook with 24 GB RAM, highlighting its performance in coding tasks.

Release of Ternary-Bonsai-2-27B-gguf, a 27B-class language model using ternary weights for extreme compression (5.9 GB) while retaining 98.2% of FP16 intelligence, optimized for efficient inference on laptops and single GPUs.

SHADOW 50M is a compact 44M-parameter AI model with ternary weights and fixed 512-bit codes, trained from scratch on 45B tokens, enabling offline inference at high speed on consumer hardware.

DeepGrove releases Maple-Preview, an open-source 20B-A1B ternary-weight reasoning LLM with SOTA reasoning for its weight class, capable of 200+ tokens/sec on a Mac mini M4 and competitive with larger models.

Presents the insight that CPU decode speed depends on active parameters per token, not total parameters, and proposes a 10B-parameter model using ternary weights and granular MoE to achieve high token rates on mid-level PCs. Reports sandbox measurements showing a speedup from 176 to 848 tok/s on an 8.3M model with minimal quality loss.

Neutrino-1 8B is an 8.19B-parameter transformer with a proprietary ternary weight format, enabling it to fit on 8GB GPUs and serve multiple platforms from a single 3.88GB artifact with high decode speed.

Prism ML releases Bonsai Image, a 1.21 GB text-to-image diffusion transformer using ternary weights (1.58-bit) for NVIDIA GPUs, offering 4.5s / 1024² on RTX 3080 and much smaller than FP16.

A highly efficient AI model architecture using ternary weights (-1, 0, 1) that achieves competitive performance while requiring only 1.58 bits per parameter, enabling deployment on extremely constrained devices.