UNAversal 8x7B V1beta GPTQ by TheBloke

 ยป  All LLMs  ยป  TheBloke  ยป  UNAversal 8x7B V1beta GPTQ   URL Share it on

  4-bit   Autotrain compatible Base model:fblgit/unaversal-8x... Base model:quantized:fblgit/un...   Conversational   En   Gptq   Juanako   Mixtral   Moe   Quantized   Region:us   Safetensors   Una

UNAversal 8x7B V1beta GPTQ Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
UNAversal 8x7B V1beta GPTQ (TheBloke/UNAversal-8x7B-v1beta-GPTQ)
๐ŸŒŸ Advertise your project ๐Ÿš€

UNAversal 8x7B V1beta GPTQ Parameters and Internals

Model Type 
mixtral
Use Cases 
Considerations:
Can be fine-tuned further, try 2e-5 or 1e-4 (since it's MOE); Can be merged, improvise and report findings
Additional Notes 
This is just a beta release; it offers high GSM/Math and TQA, recommended for merging with other mixtrals
Supported Languages 
English (full)
Training Details 
Data Sources:
LLaMA-Factory datasets
Methodology:
LLaMA-Factory datasets, example alpaca
Model Architecture:
UNA on the router_logit
Input Output 
Input Format:
{prompt}
Output Format:
text
LLM NameUNAversal 8x7B V1beta GPTQ
Repository ๐Ÿค—https://huggingface.co/TheBloke/UNAversal-8x7B-v1beta-GPTQ 
Model NameUNAversal 8X7B v1Beta
Model CreatorFBL
Base Model(s)  UNAversal 8x7B V1beta   fblgit/UNAversal-8x7B-v1beta
Model Size6.1b
Required VRAM23.8 GB
Updated2025-09-23
MaintainerTheBloke
Model Typemixtral
Model Files  23.8 GB
Supported Languagesen
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitectureMixtralForCausalLM
Licensecc-by-nc-sa-4.0
Context Length32768
Model Max Length32768
Transformers Version4.37.0.dev0
Tokenizer ClassLlamaTokenizer
Vocabulary Size32000
Torch Data Typebfloat16

Best Alternatives to UNAversal 8x7B V1beta GPTQ

Best Alternatives
Context / RAM
Downloads
Likes
...ixtral 8x7B Instruct V0.1 GPTQ32K / 23.8 GB204520138
Mixtral 8x7B V0.1 GPTQ32K / 23.8 GB1626127
...Hermes 2 Mixtral 8x7B DPO GPTQ32K / 23.8 GB210326
...tLM Mixtral 8x7B Instruct GPTQ32K / 23.8 GB13733
Dolphin 2.5 Mixtral 8x7b GPTQ32K / 23.8 GB55114
...Hermes 2 Mixtral 8x7B SFT GPTQ32K / 23.8 GB1311
Dolphin 2.7 Mixtral 8x7b GPTQ32K / 23.8 GB7819
Bagel DPO 8x7b V0.2 GPTQ32K / 23.8 GB52
Open Gpt4 8x7B V0.2 GPTQ32K / 23.8 GB96
...xtral Instruct 8x7b Zloss GPTQ32K / 23.8 GB112
Note: green Score (e.g. "73.2") means that the model is better than TheBloke/UNAversal-8x7B-v1beta-GPTQ.

Rank the UNAversal 8x7B V1beta GPTQ Capabilities

๐Ÿ†˜ Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! ๐ŸŒŸ

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  

What open-source LLMs or SLMs are you in search of? 51534 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124