What are the hardware requirements for Bloom 1b7 8bit?

Bloom 1b7 8bit requires approximately 2.2 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Bloom 1b7 8bit and how large is it?

Bloom 1b7 8bit is developed by ybelkada, a model with 1.7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Bloom 1b7 8bit?

Bloom 1b7 8bit is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Bloom 1b7 8bit by ybelkada — VRAM 2.2GB

Name: Bloom 1b7 8bit
Author: ybelkada

Bloom 1b7 8bit is an open-source language model by ybelkada. Features: 1.7b LLM, VRAM: 2.2GB, License: bigscience-bloom-rail-1.0, Quantized, LLM Explorer Score: 0.06.

Arxiv:2208.07339 8-bit 8bit Ak Ar As Bloom Bm Bn Ca Code En Endpoints compatible Es Eu Fon Fr Gu Hi Id Ig Ki Kn Lg Ln Ml Mr Ne Nso Ny Or Pa Pt Pytorch Quantized 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/ybelkada/bloom-1b7-8bit

Bloom 1b7 8bit Parameters and Internals

Model Type

text generation

Additional Notes

The model's state dict includes quantized int8 weights and float16 quantization statistics.

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 (Oriya), 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)

Release Notes

Version:

1.0

Date:

2022-05-26

Notes:

First release of the 8bit model weights.

LLM Name	Bloom 1b7 8bit
Repository 🤗	https://huggingface.co/ybelkada/bloom-1b7-8bit
Model Size	1.7b
Required VRAM	2.2 GB
Updated	2026-07-31
Maintainer	ybelkada
Model Type	bloom
Model Files	2.2 GB 2.2 GB
Supported Languages	ak 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
Quantization Type	8bit
Model Architecture	BloomForCausalLM
License	bigscience-bloom-rail-1.0
Transformers Version	4.28.0.dev0
Tokenizer Class	BloomTokenizer
Padding Token	<pad>
Vocabulary Size	250880
Torch Data Type	float16

Rank the Bloom 1b7 8bit capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

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

Best Alternatives to Bloom 1b7 8bit

Best Alternatives	Context / RAM	Downloads	Likes
Bloom Lora 8bit	0K / 2.3 GB	7	0
Merged AutoGPTQ DPO	0K / 2.3 GB	5	0
Merged DPO Model	0K / 6.8 GB	5	0
Bloom20	0K / 3.4 GB	5	0
Mnlp DPO Model Bloom	0K / 6.8 GB	7	0
Bloom 1b7	0K / 3.4 GB	19777	123
Bloom 1b7 Fp32	0K / 6.8 GB	10	0
Aira 2 Portuguese 1B7	0K / 0 GB	15	2
Bloomz 1b7	0K / 3.4 GB	934	23
Bloom 1b7 Intermediate	0K / 3.4 GB	1022	0

Note: green Score (e.g. "73.2") means that the model is better than ybelkada/bloom-1b7-8bit.