What are the hardware requirements for Dragoman?

Dragoman requires approximately 2.7 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Dragoman and how large is it?

Dragoman is developed by lang-uk, a model with 7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Dragoman?

Dragoman is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Dragoman by lang-uk — VRAM 2.7GB

Name: Dragoman
Author: lang-uk

Dragoman is an open-source language model by lang-uk. Features: 7b LLM, VRAM: 2.7GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.12.

Adapter Base model:adapter:mistralai/m... Base model:mistralai/mistral-7... Dataset:helsinki-nlp/opus para... Dataset:turuta/multi30k-uk En Finetuned Ggml Lora Model-index Peft Quantized Region:us Safetensors Translation Uk

Model Card on HF 🤗: https://huggingface.co/lang-uk/dragoman

Dragoman Parameters and Internals

Model Type

Translation model

Use Cases

Primary Use Cases:

Sentence-level English -> Ukrainian translation

Limitations:

Performance on multi-sentence texts is not guaranteed.

Supported Languages

Source Language (English), Target Language (Ukrainian)

Training Details

Data Sources:

Helsinki-NLP/opus_paracrawl, turuta/Multi30k-uk

Data Volume:

3M pairs of Ukrainian and English sentences + 17K examples for higher quality

Methodology:

Supervised finetuning with noisy parallel dataset followed by k-fold perplexity filtering.

LLM Name	Dragoman
Repository 🤗	https://huggingface.co/lang-uk/dragoman
Base Model(s)	mistralai/Mistral-7B-v0.1 mistralai/Mistral-7B-v0.1
Model Size	7b
Required VRAM	2.7 GB
Updated	2026-08-03
Maintainer	lang-uk
Model Files	3.2 GB 2.7 GB
Supported Languages	uk en
GGML Quantization	Yes
Quantization Type	ggml
Model Architecture	Adapter
License	apache-2.0
Is Biased	none
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	up_proj\|q_proj\|v_proj\|gate_proj\|o_proj\|lm_head\|down_proj\|k_proj
LoRA Alpha	128
LoRA Dropout	0.05
R Param	256

Rank the Dragoman capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Dragoman

Best Alternatives	Context / RAM	Downloads	Likes
Llama 2 Medical Consultation	0K / 0 GB	13	5
...en2.5 Coder 7B Vietnamese LoRA	0K / 0.3 GB	45	0
Atomgpt Mistral Tc Supercon	0K / 0.2 GB	12	6
...hat Hf Hotel Booking Assistant	0K / 0.2 GB	5	0
Mistral Numericnlg FV	0K / 0.3 GB	6	0
Mistral Wikitable FV	0K / 0.3 GB	6	0
Mistral Charttotext FV	0K / 0.3 GB	5	0
Gemma 7B Us Minecraft	0K / 0 GB	7	2
Gemma7b SFT	0K / 0 GB	19	0
Gemma7b SFT	0K / 0 GB	6	0

Note: green Score (e.g. "73.2") means that the model is better than lang-uk/dragoman.