LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

Nous Hermes 13B GPTQ by TheBloke

 »  All LLMs  »  TheBloke  »  Nous Hermes 13B GPTQ   URL Share it on

  762   
  4-bit   Autotrain compatible   Distillation   En   Gptq   Llama   Quantized   Region:us   Safetensors   Self-instruct
Model Card on HF 🤗: https://huggingface.co/TheBloke/Nous-Hermes-13B-GPTQ 

Nous Hermes 13B GPTQ Benchmarks

ARC: 56.57 vs 96.7 (so35)-41.5%
HellaSwag: 82.11 vs 95.3 (gpt4)-13.8%
MMLU: 50.44 vs 88.3 (so35)-42.9%
TruthfulQA: 51.5 vs 59 (gpt4)-12.7%
WinoGrande: 75.3 vs 87.5 (gpt4)-13.9%
GSM8K: 8.34 vs 96.4 (so35)-91.3%
LLME Score: 0.14587
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Nous Hermes 13B GPTQ (TheBloke/Nous-Hermes-13B-GPTQ)

Nous Hermes 13B GPTQ Parameters and Internals

Model Type 
language model, text generation
Use Cases 
Areas:
research, commercial applications
Primary Use Cases:
long response generation, low hallucination generation
Limitations:
not specified
Additional Notes 
Benchmarks are pending. Compute provided by Redmond AI.
Supported Languages 
en (high)
Training Details 
Data Sources:
GPTeacher, general roleplay v1&2, code instruct datasets, Nous Instruct & PDACTL, CodeAlpaca, Evol_Instruct Uncensored, GPT4-LLM, Unnatural Instructions, Camel-AI's Biology/Physics/Chemistry and Math Datasets, Airoboros' GPT-4 Dataset
Data Volume:
300,000 instructions
Methodology:
Fine-tuned on synthetic GPT-4 outputs; sequence length of 2000.
Context Length:
2000
Training Time:
50+ hours on an 8x a100 80GB DGX machine
Hardware Used:
8x a100 80GB DGX machine
Model Architecture:
Enhanced Llama 13b model through fine-tuning.
Input Output 
Input Format:
Alpaca prompt format
Accepted Modalities:
text
Output Format:
Textual responses
Release Notes 
Version:
GPTQ 4bit
Notes:
Quantisation to 4bit using GPTQ-for-LLaMa.
Version:
FP16
Notes:
Model uploaded in FP16 format.
LLM NameNous Hermes 13B GPTQ
Repository 🤗https://huggingface.co/TheBloke/Nous-Hermes-13B-GPTQ 
Base Model(s)  Nous Hermes 13B   NousResearch/Nous-Hermes-13b
Model Size13b
Required VRAM7.5 GB
Updated2025-05-04
MaintainerTheBloke
Model Typellama
Model Files  7.5 GB
Supported Languagesen
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitectureLlamaForCausalLM
Licenseother
Context Length2048
Model Max Length2048
Transformers Version4.29.2
Tokenizer ClassLlamaTokenizer
Beginning of Sentence Token<s>
End of Sentence Token</s>
Unk Token<unk>
Vocabulary Size32001
Torch Data Typebfloat16

Best Alternatives to Nous Hermes 13B GPTQ

Best Alternatives
Context / RAM
Downloads
Likes
Yarn Llama 2 13B 128K GPTQ128K / 7.3 GB1216
LongAlign 13B 64K GPTQ64K / 7.3 GB101
...boros L2 13B 2 1 YaRN 64K GPTQ64K / 7.3 GB143
Yarn Llama 2 13B 64K GPTQ64K / 7.3 GB81
OrcaMaid V3 13B 32K GPTQ32K / 7.3 GB83
OrcaMaid V2 FIX 13B 32K GPTQ32K / 7.3 GB74
EverythingLM 13B 16K GPTQ16K / 7.3 GB613
Tinybra 13B GPTQ 32g 4BIT16K / 8 GB91
Tinybra 13B GPTQ 4BIT16K / 7 GB110
WhiteRabbitNeo 13B GPTQ16K / 7.3 GB54
Note: green Score (e.g. "73.2") means that the model is better than TheBloke/Nous-Hermes-13B-GPTQ.

Rank the Nous Hermes 13B GPTQ Capabilities

🆘 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! 🌟

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? 46981 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227
  Support LLM Explorer