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

Bloom 1b7 by bigscience

 »  All LLMs  »  bigscience  »  Bloom 1b7   URL Share it on

  72227   
  Arxiv:1909.08053   Arxiv:2108.12409   Arxiv:2110.02861   Ak   Ar   As   Autotrain compatible   Bloom   Bm   Bn   Ca   Code   En   Endpoints compatible   Es   Eu   Fon   Fr   Gu   Hi   Id   Ig   Jax   Ki   Kn   Lg   Ln   Ml   Mr   Ne   Nso   Ny   Or   Pa   Pt   Pytorch   Region:us   Rn   Rw   Safetensors   Sn   St   Sw   Ta   Te   Tn   Ts   Tum   Tw   Ur   Vi   Wo   Xh   Yo   Zh   Zhs   Zht   Zu
Model Card on HF 🤗: https://huggingface.co/bigscience/bloom-1b7 

Bloom 1b7 Benchmarks

MMLU Pro: 0.96
GPQA: 1.12
MUSR: 6.84
BBH: 4.4
IFEval: 10.44 vs 88 (so35)-88.1%
ARC: 30.63 vs 96.7 (so35)-68.3%
HellaSwag: 47.6 vs 95.3 (gpt4)-50.1%
MMLU: 27.48 vs 88.3 (so35)-68.9%
TruthfulQA: 41.31 vs 59 (gpt4)-30%
WinoGrande: 56.04 vs 87.5 (gpt4)-36%
GSM8K: 0.83 vs 96.4 (so35)-99.1%
MATH Lvl 5: 0.08
LLME Score: 0.1443
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Bloom 1b7 (bigscience/bloom-1b7)

Bloom 1b7 Parameters and Internals

Model Type 
Transformer-based Language Model, Text Generation
Use Cases 
Areas:
Research, Exploration of language characteristics
Applications:
Text generation, Language modeling tasks
Primary Use Cases:
Language generation, Pretrained base model for task-specific fine-tuning
Limitations:
Not to be used in high-stakes settings, Biometric, political, financial domain applications
Considerations:
Models should be used with care, acknowledging bias and ensuring transparency.
Additional Notes 
Includes detailed ethical considerations for use cases and provides a comprehensive training dataset overview.
Supported Languages 
ak (Akan), ar (Arabic), as (Assamese), bm (Bambara), bn (Bengali), ca (Catalan), code (Code), en (English), es (Spanish), eu (Basque), fon (Fon), fr (French), gu (Gujarati), hi (Hindi), id (Indonesian), ig (Igbo), ki (Kikuyu), kn (Kannada), lg (Ganda), ln (Lingala), ml (Malayalam), mr (Marathi), ne (Nepali), nso (Northern Sotho), ny (Chichewa), or (Odia), pa (Punjabi), pt (Portuguese), rn (Kirundi), rw (Kinyarwanda), sn (Shona), st (Southern Sotho), sw (Swahili), ta (Tamil), te (Telugu), tn (Tswana), ts (Tsonga), tum (Tumbuka), tw (Twi), ur (Urdu), vi (Vietnamese), wo (Wolof), xh (Xhosa), yo (Yoruba), zh (Chinese), zhs (Simplified Chinese), zht (Traditional Chinese), zu (Zulu)
Training Details 
Data Sources:
45 natural languages, 12 programming languages
Data Volume:
1.5TB of pre-processed text, 350B unique tokens
Methodology:
Modified from Megatron-LM GPT2
Context Length:
2048
Training Time:
March 11th, 2022 to May 20th, 2022
Hardware Used:
Jean Zay Public Supercomputer, 64 V100 16/32GB GPUs
Model Architecture:
Decoder-only architecture with layer normalization and ALiBI positional encodings.
Input Output 
Accepted Modalities:
text
LLM NameBloom 1b7
Repository 🤗https://huggingface.co/bigscience/bloom-1b7 
Model Size1.7b
Required VRAM3.4 GB
Updated2024-12-22
Maintainerbigscience
Model Typebloom
Model Files  3.4 GB   3.4 GB
Supported Languagesak ar as bm bn ca code en es eu fr gu hi id ig ki kn lg ln ml mr ne ny or pa pt rn rw sn st sw ta te tn ts tw ur vi wo xh yo zh zu
Model ArchitectureBloomForCausalLM
Licensebigscience-bloom-rail-1.0
Transformers Version4.20.0
Tokenizer ClassBloomTokenizerFast
Padding Token<pad>
Vocabulary Size250880

Quantized Models of the Bloom 1b7

Model
Likes
Downloads
VRAM
Bloom 1b7 GPTQ 4bit G1280272 GB
Bloom 1b7 Gptq 4bit0122 GB

Best Alternatives to Bloom 1b7

Best Alternatives
Context / RAM
Downloads
Likes
Merged DPO Model0K / 6.8 GB130
Bloom200K / 3.4 GB190
Mnlp DPO Model Bloom0K / 6.8 GB120
Aira 2 Portuguese 1B70K / 0 GB822
Bloomz 1b70K / 3.4 GB212623
Bloom 1b7 Intermediate0K / 3.4 GB280
Bloom 1b7 8bit0K / 2.2 GB5606
Note: green Score (e.g. "73.2") means that the model is better than bigscience/bloom-1b7.

Rank the Bloom 1b7 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? 40066 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 v20241217
  Support LLM Explorer