LLM EXPLORER 59,420 MODELS INDEXED

Bloomz by bigscience

By bigscience · 766 downloads
HF Repo ↗ Copy URL

Bloomz is an open-source language model by bigscience. Features: 176.2b LLM, VRAM: 4.9GB, License: bigscience-bloom-rail-1.0, LLM Explorer Score: 0.05.

  Arxiv:2211.01786   Ak   Ar   As   Bloom   Bm   Bn   Ca   Code   Dataset:bigscience/xp3   Deploy:azure   En   Es   Eu   Fon   Fr   Gu   Hi   Id   Ig   Ki   Kn   Lg   Ln   Ml   Model-index   Mr   Ne   Nso   Ny   Or   Pa   Pt   Pytorch   Region:us   Rn   Rw   Safetensors   Sn   St   Sw   Ta   Te   Tensorboard   Tn   Ts   Tum   Tw   Ur   Vi   Wo   Xh   Yo   Zh   Zu
Model Card on HF 🤗: https://huggingface.co/bigscience/bloomz 

Bloomz Parameters and Internals

Model Type 
text-generation
Use Cases 
Areas:
Coreference resolution, Natural language inference, Program synthesis, Sentence completion
Applications:
Language translation, Text understanding across languages
Additional Notes 
Illustrates the potential of instruction-tuned models to perform across languages and tasks.
Supported Languages 
ak (supported), ar (supported), as (supported), bm (supported), bn (supported), ca (supported), code (supported), en (supported), es (supported), eu (supported), fon (supported), fr (supported), gu (supported), hi (supported), id (supported), ig (supported), ki (supported), kn (supported), lg (supported), ln (supported), ml (supported), mr (supported), ne (supported), nso (supported), ny (supported), or (supported), pa (supported), pt (supported), rn (supported), rw (supported), sn (supported), st (supported), sw (supported), ta (supported), te (supported), tn (supported), ts (supported), tum (supported), tw (supported), ur (supported), vi (supported), wo (supported), xh (supported), yo (supported), zh (supported), zu (supported)
Training Details 
Data Sources:
datasets: bigscience/xP3
Data Volume:
2.09 billion tokens
Methodology:
Multitask finetuning on xP3mt and zero-shot crosslingual generalization.
Hardware Used:
288 A100 80GB GPUs with 8 GPUs per node (36 nodes) using NVLink 4 inter-gpu connects, 4 OmniPath links
Model Architecture:
The same as BLOOM, refer to the config.json file.
Input Output 
Input Format:
Text prompt in any of the supported languages or formats.
Accepted Modalities:
text
Output Format:
Generated text in the same or different languages as input.
Performance Tips:
Better performance with clear termination in prompts and sufficient context.
LLM NameBloomz
Repository 🤗https://huggingface.co/bigscience/bloomz 
Model Size176.2b
Required VRAM4.9 GB
Updated2026-07-23
Maintainerbigscience
Model Typebloom
Model Files  7.2 GB: 1-of-72   4.9 GB: 2-of-72   4.9 GB: 3-of-72   4.9 GB: 4-of-72   4.9 GB: 5-of-72   4.9 GB: 6-of-72   4.9 GB: 7-of-72   4.9 GB: 8-of-72   4.9 GB: 9-of-72   4.9 GB: 10-of-72   4.9 GB: 11-of-72   4.9 GB: 12-of-72   4.9 GB: 13-of-72   4.9 GB: 14-of-72   4.9 GB: 15-of-72   4.9 GB: 16-of-72   4.9 GB: 17-of-72   4.9 GB: 18-of-72   4.9 GB: 19-of-72   4.9 GB: 20-of-72   4.9 GB: 21-of-72   4.9 GB: 22-of-72   4.9 GB: 23-of-72   4.9 GB: 24-of-72   4.9 GB: 25-of-72   4.9 GB: 26-of-72   4.9 GB: 27-of-72   4.9 GB: 28-of-72   4.9 GB: 29-of-72   4.9 GB: 30-of-72   4.9 GB: 31-of-72   4.9 GB: 32-of-72   4.9 GB: 33-of-72   4.9 GB: 34-of-72   4.9 GB: 35-of-72   4.9 GB: 36-of-72   4.9 GB: 37-of-72   4.9 GB: 38-of-72   4.9 GB: 39-of-72   4.9 GB: 40-of-72   4.9 GB: 41-of-72   4.9 GB: 42-of-72   4.9 GB: 43-of-72   4.9 GB: 44-of-72   4.9 GB: 45-of-72
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 Version4.21.0
Vocabulary Size250880

Best Alternatives to Bloomz

Best Alternatives
Context / RAM
Downloads
Likes
Bloom0K / 4.9 GB109995033
Bloomz Mt0K / 4.9 GB10032
Bloomz P30K / 4.9 GB13624
Note: green Score (e.g. "73.2") means that the model is better than bigscience/bloomz.