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

Lamma3merge2 15B MoE by allknowingroger

 »  All LLMs  »  allknowingroger  »  Lamma3merge2 15B MoE   URL Share it on

  10   
  Autotrain compatible Base model:birgermoell/llama-3... Base model:finetune:birgermoel...   Birgermoell/llama-3-dare ties   Conversational   Endpoints compatible   Frankenmoe   Instruct   Lazymergekit   Merge   Mergekit   Mixtral   Moe Mohsenfayyaz/meta-llama-3-8b-i...   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/allknowingroger/Lamma3merge2-15B-MoE 

Lamma3merge2 15B MoE Benchmarks

ARC: 59.98 vs 96.7 (so35)-38%
HellaSwag: 79.53 vs 95.3 (gpt4)-16.5%
MMLU: 67 vs 88.3 (so35)-24.1%
TruthfulQA: 51.67 vs 59 (gpt4)-12.4%
WinoGrande: 75.77 vs 87.5 (gpt4)-13.4%
GSM8K: 67.85 vs 96.4 (so35)-29.6%
LLME Score: 0.1958
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Lamma3merge2 15B MoE (allknowingroger/Lamma3merge2-15B-MoE)

Lamma3merge2 15B MoE Parameters and Internals

Model Type 
Mixture of Experts (MoE)
Additional Notes 
Lamma3merge2-15B-MoE is a Mixture of Experts model created using LazyMergekit, combining two separate models.
LLM NameLamma3merge2 15B MoE
Repository 🤗https://huggingface.co/allknowingroger/Lamma3merge2-15B-MoE 
Base Model(s)  mohsenfayyaz/Meta-Llama-3-8B-Instruct_esnli_5000_1ep   Llama 3 Dare Ties   mohsenfayyaz/Meta-Llama-3-8B-Instruct_esnli_5000_1ep   birgermoell/Llama-3-dare_ties
Model Size8b
Required VRAM27.5 GB
Updated2024-12-03
Maintainerallknowingroger
Model Typemixtral
Instruction-BasedYes
Model Files  1.1 GB: 1-of-15   2.0 GB: 2-of-15   2.0 GB: 3-of-15   2.0 GB: 4-of-15   2.0 GB: 5-of-15   2.0 GB: 6-of-15   2.0 GB: 7-of-15   2.0 GB: 8-of-15   2.0 GB: 9-of-15   2.0 GB: 10-of-15   2.0 GB: 11-of-15   2.0 GB: 12-of-15   2.0 GB: 13-of-15   2.0 GB: 14-of-15   0.4 GB: 15-of-15
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length8192
Model Max Length8192
Transformers Version4.40.0
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<|begin_of_text|>
Vocabulary Size128256
Torch Data Typebfloat16

Best Alternatives to Lamma3merge2 15B MoE

Best Alternatives
Context / RAM
Downloads
Likes
Lamma3merge3 15B MoE8K / 27.5 GB111
Mergkit 18K / 22.6 GB50
Llama3merge8 15B MoE8K / 27.5 GB50
Note: green Score (e.g. "73.2") means that the model is better than allknowingroger/Lamma3merge2-15B-MoE.

Rank the Lamma3merge2 15B MoE 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? 42577 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