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

Falcon 40B by tiiuae

 »  All LLMs  »  tiiuae  »  Falcon 40B   URL Share it on

  131380   
  Arxiv:1911.02150   Arxiv:2005.14165   Arxiv:2101.00027   Arxiv:2104.09864   Arxiv:2205.14135   Arxiv:2306.01116   Autotrain compatible   Custom code Dataset:tiiuae/falcon-refinedw...   De   En   Es   Falcon   Fr   Pytorch   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/tiiuae/falcon-40b 

Falcon 40B Benchmarks

MMLU Pro: 16.72
GPQA: 3.13
MUSR: 5.19
BBH: 16.58
IFEval: 24.96 vs 88 (so35)-71.6%
ARC: 61.86 vs 96.7 (so35)-36%
HellaSwag: 85.28 vs 95.3 (gpt4)-10.5%
MMLU: 56.89 vs 88.3 (so35)-35.6%
TruthfulQA: 41.65 vs 59 (gpt4)-29.4%
WinoGrande: 81.29 vs 87.5 (gpt4)-7.1%
GSM8K: 21.46 vs 96.4 (so35)-77.7%
MATH Lvl 5: 1.59
LLME Score: 0.3295
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Falcon 40B (tiiuae/falcon-40b)

Falcon 40B Parameters and Internals

Model Type 
Causal decoder-only
Use Cases 
Areas:
Research on large language models
Applications:
Summarization, Text generation, Chatbot
Limitations:
Model has limited proficiency in languages other than English, German, Spanish, French
Considerations:
Finetuning and studying stereotypes and biases before production usage is recommended.
Additional Notes 
A smaller model, Falcon-7B, is also available.
Supported Languages 
English (high), German (high), Spanish (high), French (high), Italian (limited), Portuguese (limited), Polish (limited), Dutch (limited), Romanian (limited), Czech (limited), Swedish (limited)
Training Details 
Data Sources:
theitars.com/falcon-refinedweb
Data Volume:
1,000B tokens
Methodology:
Trained using FlashAttention and multiquery attention mechanisms
Context Length:
2048
Training Time:
two months
Hardware Used:
384 A100 40GB GPUs
Model Architecture:
Causal decoder-only model with FlashAttention, multiquery mechanism, and rotary position embeddings
Responsible Ai Considerations 
Fairness:
Model carries stereotypes and biases commonly encountered online
Mitigation Strategies:
Further finetuning for specific tasks
LLM NameFalcon 40B
Repository 🤗https://huggingface.co/tiiuae/falcon-40b 
Model Size40b
Required VRAM83.6 GB
Updated2025-01-20
Maintainertiiuae
Model Typefalcon
Model Files  9.5 GB: 1-of-9   9.5 GB: 2-of-9   9.5 GB: 3-of-9   9.5 GB: 4-of-9   9.5 GB: 5-of-9   9.5 GB: 6-of-9   9.5 GB: 7-of-9   9.5 GB: 8-of-9   7.6 GB: 9-of-9   9.5 GB: 1-of-9   9.5 GB: 2-of-9   9.5 GB: 3-of-9   9.5 GB: 4-of-9   9.5 GB: 5-of-9   9.5 GB: 6-of-9   9.5 GB: 7-of-9   9.5 GB: 8-of-9   7.6 GB: 9-of-9
Supported Languagesen de es fr
Model ArchitectureFalconForCausalLM
Licenseapache-2.0
Model Max Length2048
Transformers Version4.27.4
Is Biased0
Tokenizer ClassPreTrainedTokenizerFast
Vocabulary Size65024
Torch Data Typebfloat16

Best Alternatives to Falcon 40B

Best Alternatives
Context / RAM
Downloads
Likes
Tiny Random Falcon 40B2K / 0 GB1025990
... Falcon 40B Instruct 4 Bit Bnb2K / 23.9 GB190
Openbuddy Falcon 40B V16.1 4K2K / 82.6 GB7331
Falcon 40B Instruct0K / 83.6 GB1103611175
ReluFalcon 40B0K / 167.1 GB1654
Tiny Random Falcon 40B0K / 0.2 GB7250
Falcon 40B Megacode20K / 82.5 GB191
...lcon 40B Ft Alpaca Dolly Dutch0K / 82.5 GB294
...slessMegaCoder Falcon 40B Mini0K / 82.5 GB7302
Falcon 40B Megacode2 Oasst0K / 82.5 GB286

Rank the Falcon 40B 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? 41636 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