A high-quality 8-bit (Q8_0) quantization of Microsoft's Phi-1.5 (1.3B). Optimized for local coding assistance and logical reasoning with a tiny 1.8GB memory footprint. Features custom stop tokens to prevent base-model hallucinations.
Phi-2 is a Transformer with 2.7 billion parameters. It was trained using the same data sources as Phi-1.5, augmented with a new data source that consists of various NLP synthetic texts and filtered websites (source: Microsoft).
The language model Phi-1.5 is a Transformer with 1.3 billion parameters.
A 7B and 15B uncensored variant of the Dolphin model family that excels at coding, based on StarCoder2.
Śruti — a guide to Hindu / Sanātana / Vedāntic philosophy (Vedas, Upaniṣads, Gītā, darśanas). Qwen2.5-1.5B fine-tune; answers in English with Sanskrit quotations (Devanāgarī + IAST).
Perfect censorship bypass (5.0/5) + superior tool use (5.0/5). 13.8/15 total. Best-in-class for shell commands, API calls, Docker, K8s. 5GB, 16K context. Apache 2.0.
Phi-3 128K version of Q5
Unofficial Llama-3-Typhoon-1.5-8B-instruct Q4_K_M Quantization (https://huggingface.co/scb10x/llama-3-typhoon-v1.5-8b-instruct)
Unofficial OpenThaiGPT 1.5-7B-Instruct Q4_K_M Quantization (https://huggingface.co/openthaigpt/openthaigpt1.5-7b-instruct)
Unofficial OpenThaiGPT 1.5-14B-Instruct Q4_K_M Quantization (https://huggingface.co/openthaigpt/openthaigpt1.5-14b-instruct)
(UNOFFICIAL) Ported from https://huggingface.co/OuteAI/Lite-Mistral-150M-v2-Instruct-GGUF
GGUF on HF from crusoeai/dolphin-2.9.3-qwen2-1.5b-GGUF
dolphin-2.9.3-Yi-1.5-34B-32k-Q5_K_M
Convert and Quantize from https://huggingface.co/cognitivecomputations/dolphin-2.9.1-yi-1.5-9b
Razonar: dolphin-yi-1.5-34b