Needle 2 14MB AI: 1500 TPS Offline AI Router Just Dropped! - On-device structured expense parser & local Streamlit dashboard.
-
Updated
Aug 16, 2026 - Python
#
needle2
Here are 7 public repositories matching this topic...
Open-source CPU inference and PyTorch tools for Needle 2: CQ2/CQ4 quantization, model conversion, QAT, and ARM NEON/SDOT kernels.
pytorch arm64 quantization quantization-aware-training model-conversion cpu-inference function-calling needle2
-
Updated
Sep 9, 2026 - Python
Chinese fine-tuning and Raspberry Pi 5 deployment for Needle 2 tool calling
-
Updated
Aug 31, 2026 - Python
Golang rewrite of Cactus-Compute needle2 library
-
Updated
Sep 11, 2026 - Go
Run Cactus Needle 2 on Android 5.1 / ARMv7 hardware by relinking the official ARMv7 library with Android 5 compatibility fixes. Tested on a Lenovo A2020A40 with successful native inference.
android armv7 android-ndk bionic edge-ai on-device-ai old-hardware legacy-android android5 tool-calling native-inference needle2
-
Updated
Sep 15, 2026 - Shell
Thai synthetic tool-calling dataset + multi-speaker Thai TTS pipeline (FastThaiG2P × Kokoro-82M) + LoRA fine-tuning scripts for the Needle 2 edge LLM (45M params)
text-to-speech tts speech-synthesis dataset needle thai lora thai-nlp voice-assistant synthetic-data fine-tuning kokoro jax edge-ai llm function-calling small-language-model tool-calling needle2 fastthaig2p
-
Updated
Aug 20, 2026 - Python
Add this topic to your repo
To associate your repository with the needle2 topic, visit your repo's landing page and select "manage topics."