LLM EXPLORER 60,519 MODELS INDEXED

Mistral 7B V0.3 Chinese Chat 4bit by shenzhi-wang

By shenzhi-wang · 45 downloads
HF Repo ↗ Copy URL

Mistral 7B V0.3 Chinese Chat 4bit is an open-source language model by shenzhi-wang. Features: 7b LLM, VRAM: 4.1GB, Context: 32K, License: apache-2.0, Quantized, Instruction-Based, LLM Explorer Score: 0.12.

  Arxiv:2403.07691   4bit Base model:mistralai/mistral-7... Base model:quantized:mistralai...   Conversational   En   Endpoints compatible   Gguf   Instruct   Llama-factory   Mistral   Orpo   Q4   Quantized   Region:us   Zh
Model Card on HF 🤗: https://huggingface.co/shenzhi-wang/Mistral-7B-v0.3-Chinese-Chat-4bit 

Mistral 7B V0.3 Chinese Chat 4bit Parameters and Internals

Model Type 
text generation, instruction-tuned
Use Cases 
Areas:
Chinese and English user applications
Applications:
roleplaying, tool use, mathematics
Primary Use Cases:
text generation in Chinese and English, chat and instruction following
Additional Notes 
The model significantly improves Chinese language capabilities compared to its base model. Uses meticulously fine-tuned ORPO methodology to handle multilingual interactions efficiently.
Supported Languages 
en (fluent), zh (fluent)
Training Details 
Data Sources:
Chinese-English dataset of ~100K preference pairs
Data Volume:
~100K preference pairs
Methodology:
Full-parameter fine-tuning using ORPO
Context Length:
32768
Model Architecture:
Built on Mistral-7B-Instruct-v0.3
Input Output 
Input Format:
Text input in form of user-system interaction
Accepted Modalities:
text
Output Format:
Textual responses, including roleplaying and conversation prompts
Performance Tips:
Avoid identity-related questions as they may yield random responses
Release Notes 
Version:
v0.3
Date:
May 26, 2024
Notes:
First model fine-tuned for Chinese and English users based on Mistral-7B-Instruct-v0.3
LLM NameMistral 7B V0.3 Chinese Chat 4bit
Repository 🤗https://huggingface.co/shenzhi-wang/Mistral-7B-v0.3-Chinese-Chat-4bit 
Base Model(s)  mistralai/Mistral-7B-Instruct-v0.3   mistralai/Mistral-7B-Instruct-v0.3
Model Size7b
Required VRAM4.1 GB
Updated2026-08-01
Maintainershenzhi-wang
Model Typemistral
Instruction-BasedYes
Model Files  4.1 GB
Supported Languagesen zh
GGUF QuantizationYes
Quantization Typegguf|q4|4bit
Model ArchitectureMistralForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.40.0
Vocabulary Size32768
Torch Data Typebfloat16

Best Alternatives to Mistral 7B V0.3 Chinese Chat 4bit

Best Alternatives
Context / RAM
Downloads
Likes
Mistral 7B Instruct V0.2.gguf32K / 14.5 GB173
Mistral 7B Instruct V0.2 GGUF32K / 4.4 GB262
...andle Mistral 7B Instruct V0.232K / 14.4 GB2010
Tamil Mistral 7B Instruct V0.132K / 14.8 GB35014
Mistral V0.3 7B Cybersecurity32K / 14.5 GB521
BioMistralMerged32K / 14.4 GB542
...andle Mistral 7B Instruct V0.132K / 14.4 GB12363
Mistral 7B Instruct V0.3 GGUF32K / 2.7 GB1891
Mistral 7B Instruct V0.3 GGUF32K / 2.7 GB1561
...stral 7B V0.3 Chinese Chat F1632K / 14.5 GB511
Note: green Score (e.g. "73.2") means that the model is better than shenzhi-wang/Mistral-7B-v0.3-Chinese-Chat-4bit.