LLM EXPLORER 58,178 MODELS INDEXED

OmniLing V1 12B Experimental by WoonaAI

By WoonaAI · 6 downloads
HF Repo ↗ Copy URL

OmniLing V1 12B Experimental is an open-source language model by WoonaAI. Features: 12b LLM, VRAM: 24.5GB, Context: 1000K, License: cc-by-nc-4.0, Instruction-Based, LLM Explorer Score: 0.14.

Base model:finetune:mistralai/... Base model:mistralai/mistral-n...   Conversational   Dataset:woonaai/rus-eng-data   En   Endpoints compatible   Instruct   Mistral   Region:us   Ru   Safetensors   Sharded   Tensorflow   Unsloth
Model Card on HF 🤗: https://huggingface.co/WoonaAI/OmniLing-V1-12b-experimental 

OmniLing V1 12B Experimental Parameters and Internals

Model Type 
text-generation
Use Cases 
Areas:
research, non-commercial applications
Applications:
translation between Russian and English
Primary Use Cases:
translation of literary texts, datasets, user messages, descriptions
Additional Notes 
The model is experimental and stable performance is not guaranteed.
Supported Languages 
ru (Russian), en (English)
Training Details 
Data Sources:
WoonaAI/Rus-Eng-Data
Methodology:
finetuning on a stronger model's generated data
Model Architecture:
Mistral Nemo 12b
Input Output 
Input Format:
User message should contain the text to be translated (based on training format).
Accepted Modalities:
text
Output Format:
Translated text
Performance Tips:
Recommended low temperature settings T=0.1-0.3.
LLM NameOmniLing V1 12B Experimental
Repository 🤗https://huggingface.co/WoonaAI/OmniLing-V1-12b-experimental 
Base Model(s)  Mistral Nemo Instruct 2407   mistralai/Mistral-Nemo-Instruct-2407
Model Size12b
Required VRAM24.5 GB
Updated2026-07-27
MaintainerWoonaAI
Model Typemistral
Instruction-BasedYes
Model Files  4.9 GB: 1-of-5   4.9 GB: 2-of-5   4.9 GB: 3-of-5   4.9 GB: 4-of-5   4.9 GB: 5-of-5
Supported Languagesru en
Model ArchitectureMistralForCausalLM
Licensecc-by-nc-4.0
Context Length1024000
Model Max Length1024000
Transformers Version4.44.2
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<pad>
Vocabulary Size131072
Torch Data Typebfloat16

Best Alternatives to OmniLing V1 12B Experimental

Best Alternatives
Context / RAM
Downloads
Likes
...r Nemo 12B Instruct R 21 09 241000K / 24.5 GB396277141
Qliphoth 12B V1.11000K / 24.5 GB223
Qliphoth 12B V11000K / 24.5 GB262
SauerkrautLM Nemo 12B Instruct1000K / 24.5 GB147322
Magnum V4 12B1000K / 24.5 GB111654
MN Slush1000K / 24.5 GB18132
ChatWaifu V1.41000K / 24.5 GB33520
ChatWaifu 12B V2.01000K / 24.5 GB17522
Mistral Nemo Bophades 12B1000K / 24.5 GB189
Impish Nemo 12B1000K / 24.5 GB35584
Note: green Score (e.g. "73.2") means that the model is better than WoonaAI/OmniLing-V1-12b-experimental.