timidlly/
unified-v1:latest

19 2 months ago

We create refined, balanced, and multi-modal language models optimized for reliability, clean output, and contextual awareness.

2 months ago

f30719191216 · 4.7GB ·

model
llama
·
8.03B
·
Q4_0
template
{{ if .System }}<|start_header_id|>system<|end_header_id|> {{ .System }}<|eot_id|>{{ end }}{{ if .Pr
license
META LLAMA 3 COMMUNITY LICENSE AGREEMENT Meta Llama 3 Version Release Date: April 18, 2024 “Agreem
system
timidlly is a unified assistant combining the reasoning power of LLaMA 3 and the programming experti
license
timidlly
params
{ "num_keep": 24, "stop": [ "<|start_header_id|>", "<|end_header_id|>",

Readme

🌿 timidlly / unified-v1

🧬 Bio

Welcome to Timidlly — a research-focused initiative building hybrid models that blend linguistic reasoning and algorithmic precision.
We create refined, balanced, and multi-modal language models optimized for reliability, clean output, and contextual awareness.

🤖 Model: unified-v1

unified-v1 is a 4.7GB lightweight hybrid model that merges LLaMA 3’s conversational intelligence with DeepSeek’s structured reasoning.
Built for creators, developers, and AI researchers, it excels in: - Technical QA - Thoughtful explanations - Clean code generation - Creative synthesis

⚙️ Technical Overview

Feature Value
Model Size 4.7 GB
Context Length 8,000 tokens
Input Type Text
Base Models LLaMA 3 + DeepSeek mix
Interface Ollama

Run with: “`bash ollama run timidlly/unified-v1