lmg-anon / translation updated 1 year ago

vntl-llama3-8b-v2-gguf

Name: vntl-llama3-8b-v2-gguf
Author: lmg-anon

This is a LLaMA 3 Youko qlora fine-tune, created using a new version of the VNTL dataset. The purpose of this fine-tune is to improve performance of LLMs at translating Japanese visual novels to English.

Params

—

Context

—

Downloads 30d

774 K

Likes

Commercial use: conditional · llama3 Not gated GGUF 2 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

782 K757 K

Jul 28Jul 31Aug 3Aug 6

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
vntl-llama3-8b-v2-hf-q5_k_m.gguf	Q5_K_M	5.7 GB	6.8 GB	✅ Runs comfortably
vntl-llama3-8b-v2-hf-q8_0.gguf	Q8_0	8.5 GB	9.9 GB	✅ Runs comfortably

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Specifications

Licence: llama3
First seen on the Hub: 2025-01-02
Training datasets: lmg-anon/VNTL-v5-1k
Added to our catalog: 2026-07-28

Compare with

t5-small 24.7 M t5-base 1.6 M nllb-200-distilled-600M 1.6 M opus-mt-nl-en 1.2 M opus-mt-fr-en 900 K opus-mt-en-ru 686 K

Appears in

Best open coding models Commercial-use text models Models that fit in 12 GB