What are the hardware requirements for Bloomz 3B Sft Chat?

Bloomz 3B Sft Chat requires approximately 6 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Bloomz 3B Sft Chat and how large is it?

Bloomz 3B Sft Chat is developed by cmarkea, a model with 3b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Bloomz 3B Sft Chat perform on standard benchmarks?

Bloomz 3B Sft Chat has the following published scores: MMLU 31.49. Compare against reference models on this page or on the LLM Explorer leaderboards.

Bloomz 3B Sft Chat by cmarkea — VRAM 6GB

Name: Bloomz 3B Sft Chat
Author: cmarkea

Bloomz 3B Sft Chat is an open-source language model by cmarkea. Features: 3b LLM, VRAM: 6GB, License: bigscience-bloom-rail-1.0, LLM Explorer Score: 0.13, Arc: 36.9, HellaSwag: 54.3, MMLU: 31.5, GSM8K: 0.4.

Arxiv:2001.09977 Arxiv:2012.15613 Bloom Dataset:ehartford/wizard vicun... Dataset:laion/oig Dataset:shahules786/orca-chat Dataset:timdettmers/openassist... En Endpoints compatible Fr Pytorch Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/cmarkea/bloomz-3b-sft-chat

Bloomz 3B Sft Chat Parameters and Internals

Model Type

text generation, chatbot

Use Cases

Areas:

chatbot context

Primary Use Cases:

zero-shot use cases

Limitations:

Not suitable for languages other than English and French

Additional Notes

Performance gains are logarithmic in relation to the increase in model parameters. The change from float16 to bfloat16 may degrade performance in languages other than French and English.

Supported Languages

fr (high), en (high)

Training Details

Data Sources:

ehartford/wizard_vicuna_70k_unfiltered, shahules786/orca-chat, timdettmers/openassistant-guanaco, laion/OIG

Data Volume:

0.9 billion tokens

Hardware Used:

1 x A100 40GB

Input Output

Input Format:

[human prompt] [bot answer]

Output Format:

text

Performance Tips:

Precede prompt with BOS token and generated part with EOS token.

LLM Name	Bloomz 3B Sft Chat
Repository 🤗	https://huggingface.co/cmarkea/bloomz-3b-sft-chat
Model Size	3b
Required VRAM	6 GB
Updated	2026-07-29
Maintainer	cmarkea
Model Type	bloom
Model Files	6.0 GB 6.0 GB
Supported Languages	fr en
Model Architecture	BloomForCausalLM
License	bigscience-bloom-rail-1.0
Transformers Version	4.31.0
Tokenizer Class	BloomTokenizer
Padding Token	<pad>
Vocabulary Size	250880
Torch Data Type	bfloat16

Rank the Bloomz 3B Sft Chat 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 Bloomz 3B Sft Chat

Best Alternatives	Context / RAM	Downloads	Likes
Bloom 3B Conversational	0K / 6.1 GB	34	2
Blossom V2 3B	0K / 6 GB	56	0
Gpt3 Finnish 3B Instruct	0K / 12.7 GB	4	1
... Bloom 3B Conversational 4bits	0K / 2.6 GB	10	0
Blossom V1 3B	0K / 6 GB	54	1
Bloom Zh 3B Chat	0K / 6 GB	1925	11
Gogpt 3B Bloom	0K / 14.6 GB	69	5
Deer 3B	0K / 14.6 GB	63	2
Bloom 3B	0K / 6 GB	17625	94
...lish Dental Raw New Pretrained	0K / 3.6 GB	6	0

Note: green Score (e.g. "73.2") means that the model is better than cmarkea/bloomz-3b-sft-chat.