LLM EXPLORER 61,561 MODELS INDEXED

Leo Hessianai 7B Chat Bilingual by LeoLM

By LeoLM · 118 downloads
HF Repo ↗ Copy URL

Leo Hessianai 7B Chat Bilingual is an open-source language model by LeoLM. Features: 7b LLM, VRAM: 13.5GB, Context: 8K, Instruction-Based, LLM Explorer Score: 0.12, Arc: 51, HellaSwag: 76, MMLU: 44.7, GSM8K: 2.7.

  Custom code Dataset:bjoernp/oasst25-08-23-... Dataset:freedomintelligence/al... Dataset:freedomintelligence/ev... Dataset:garage-baind/open-plat...   Dataset:leolm/german poems   Dataset:leolm/german songs   Dataset:leolm/openschnabeltier   Dataset:openassistant/oasst-de Dataset:wizardlm/wizardlm evol...   De   En   Endpoints compatible   Instruct   Llama   Pytorch   Region:us   Sharded
Model Card on HF 🤗: https://huggingface.co/LeoLM/leo-hessianai-7b-chat-bilingual 

Leo Hessianai 7B Chat Bilingual Parameters and Internals

Model Type 
Causal decoder-only transformer language model
Use Cases 
Applications:
Writing, Explanation, Discussion tasks
Primary Use Cases:
Bilingual chat in English and German
Limitations:
Struggles with math and advanced reasoning
Considerations:
Developers should perform safety testing prior to deployment.
Additional Notes 
Developed in collaboration with LAION and HessianAI.
Supported Languages 
en (Bilingual support in English and German), de (Bilingual support in English and German)
Training Details 
Data Sources:
LeoLM/OpenSchnabeltier, OpenAssistant/OASST-DE, FreedomIntelligence/alpaca-gpt4-deutsch, FreedomIntelligence/evol-instruct-deutsch, LeoLM/German_Poems, LeoLM/German_Songs, garage-bAInd/Open-Platypus, WizardLM/WizardLM_evol_instruct_70k, bjoernp/oasst25-08-23-filtered
Data Volume:
Total Accepted: 235632 samples Total Accepted tokens: 115,862,397
Methodology:
Continued pretraining on a large corpus of German-language text
Context Length:
8000
Hardware Used:
HessianAI's new supercomputer 42
Safety Evaluation 
Ethical Considerations:
Potential outputs cannot be predicted in advance, may include inaccurate, biased, or objectionable responses.
Input Output 
Input Format:
ChatML format
Accepted Modalities:
text
Output Format:
text
Performance Tips:
Can utilize flash-attn2 for faster inference.
LLM NameLeo Hessianai 7B Chat Bilingual
Repository 🤗https://huggingface.co/LeoLM/leo-hessianai-7b-chat-bilingual 
Model Size7b
Required VRAM13.5 GB
Updated2026-07-31
MaintainerLeoLM
Model Typellama
Instruction-BasedYes
Model Files  10.0 GB: 1-of-2   3.5 GB: 2-of-2
Supported Languagesen de
Model ArchitectureLlamaForCausalLM
Context Length8192
Model Max Length8192
Transformers Version4.33.1
Tokenizer ClassLlamaTokenizer
Beginning of Sentence Token<s>
End of Sentence Token</s>
Unk Token<unk>
Vocabulary Size32128
Torch Data Typefloat16

Quantized Models of the Leo Hessianai 7B Chat Bilingual

Model
Likes
Downloads
VRAM
...ssianai 7B Chat Bilingual GGUF47852 GB
...essianai 7B Chat Bilingual AWQ2153 GB
...ssianai 7B Chat Bilingual GPTQ063 GB

Best Alternatives to Leo Hessianai 7B Chat Bilingual

Best Alternatives
Context / RAM
Downloads
Likes
1241024K / 16.1 GB930
A5.41024K / 16.1 GB120
A3.41024K / 16.1 GB130
A2.41024K / 16.1 GB120
... Qwen2.5llamaify 7B V23.1 200K195K / 15.2 GB295
Salamandra 7B Instruct 2606160K / 15.6 GB8173
SuperNeuralDreadDevil 8B128K / 16.1 GB271
Falcon3 7B Instruct32K / 14.8 GB1174980
Falcon3 Jessi V0.4 7B Slerp32K / 14.9 GB349
Jessi V0.4 Falcon3 7B Instruct32K / 14.8 GB220
Note: green Score (e.g. "73.2") means that the model is better than LeoLM/leo-hessianai-7b-chat-bilingual.