LFM3-1.7B is a tri-hybrid model for local agents and specialized applications.
SmolLM2 is a family of compact language models available in three size: 135M, 360M, and 1.7B parameters.
Fine-tuned version of LiquidAI/LFM3-1.7B aligned with Direct Preference Optimization (DPO) to eliminate hallucinations in Retrieval-Augmented Generation (RAG) pipelines. Small language models (< 3B) often fail to admit when retrieved documents do not con
pulled as BF16, vocab slice (combined_en_huge) requant to IQ4_NL (combined_en_huge) -- (English + Coding focus) - ~20% faster than Q4_K_M
pulled as BF16, requant to IQ4_NL (combined_en_huge) -- (English + Coding focus)
LFM2-8B-A1B is the best on-device MoE in terms of both quality (comparable to 3-4B dense models) and speed (faster than Qwen3-1.7B). Updated LFM2.5-8B-A1B is now available!
LFM2.5-2.6B by Liquid AI — deploy agents everywhere. A 2.6B dense reasoning model (Q4_K_M, ~1.7 GB) with native tool calling, 128K context, and 16-language support. Runs on any 8 GB GPU.
Copy of https://huggingface.co/allenai/OLMoE-1B-7B-0125-Instruct-GGUF
Unofficial THaLLE-0.1-7B-fa Q4_K_M Quantization (https://huggingface.co/KBTG-Labs/THaLLE-0.1-7B-fa)
Llama-3.1-SauerkrautLM-70b-Instruct is a fine-tuned Model based on meta-llama/Meta-Llama-3.1-70B-Instruct.
A Mixture-of-Experts LLM with 1B active and 7B total parameters.