librepowerai/TriLM-1.5B-TQ2_0

TriLM 1.5B — TQ2_0 GGUF for IBM Power

TriLM 1.5B (Spectra Suite, ternary {-1, 0, +1} weights) quantized to TQ2_0 (2.06 bits per weight) for fast CPU inference with llama.cpp — optimized for IBM POWER9 and later with the LibrePower VSX ternary kernels.

• File: TriLM-1.5B-TQ2_0.gguf (716 MiB, 1.52 B parameters)
• Quantization: TQ2_0 — exact ternary, no quality loss vs the original ternary weights
• No GPU required.

Run it on IBM Power (ppc64le)

curl -fsSL https://linux.librepower.org/install.sh | sudo sh
sudo apt install librepower-llama

lp-llama-completion -m TriLM-1.5B-TQ2_0.gguf -p "Once upon a time" -n 64 -t 16

Works with any recent llama.cpp on any architecture; the LibrePower build adds VSX acceleration on Power (4.3x prompt / 2.4x generation vs the generic path).

AIX / big-endian

TriLM-1.5B-TQ2_0-be.gguf is the big-endian variant for IBM AIX (dnf install llama-aix from aix.librepower.org):

lp-llama-completion -m TriLM-1.5B-TQ2_0-be.gguf -p "Once upon a time" -n 64 -t 16

Performance (IBM POWER9, Ubuntu 22.04, llama-bench)

Test	Threads	Tokens/s
Prompt processing (pp64)	16	122.1
Generation (tg32)	16	33.7

Note: this is a base model (no instruction tuning) — use completion-style prompts, not chat.

Credits

• Model: Spectra Suite (Apache 2.0)
• TQ2_0 quantization & Power packaging: LibrePower