What are the hardware requirements for Normistral 7B Warm?

Normistral 7B Warm requires approximately 14.5 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Normistral 7B Warm and how large is it?

Normistral 7B Warm is developed by norallm, a model with 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 Normistral 7B Warm?

Normistral 7B Warm is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Normistral 7B Warm by norallm — VRAM 14.5GB, 2K context

Name: Normistral 7B Warm
Author: norallm

Normistral 7B Warm is an open-source language model by norallm. Features: 7b LLM, VRAM: 14.5GB, Context: 2K, License: apache-2.0, Quantized, LLM Explorer Score: 0.13.

Arxiv:2005.14165 Arxiv:2204.02311 Arxiv:2302.01398 Dataset:nbailab/ncc Dataset:uonlp/culturax Dataset:vikp/starcoder filtere... Deploy:azure Endpoints compatible Generative Gguf Gpt Mistral Nb Nn No Pytorch Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/norallm/normistral-7b-warm

Normistral 7B Warm Parameters and Internals

Model Type

auto-regressive, text generation

Additional Notes

The model is primarily intended for research purposes, and it is pretrained on open Norwegian data. It can generate harmful completions if prompted inappropriately.

Supported Languages

no (high), nb (high), nn (high)

Training Details

Data Sources:

NCC corpus, HPLT corpus, CulturaX, Starcoder

Data Volume:

260 billion tokens

Methodology:

Warm-started training from Mistral-7b-v0.1 architecture.

Context Length:

2000

Training Time:

December 2023 to January 2024

Hardware Used:

LUMI cluster in Finland

Model Architecture:

optimized transformer architecture based on the Mistral/Llama language models

Input Output

Input Format:

Textual input

Accepted Modalities:

text

Output Format:

Generated text

Release Notes

Version:

Initial

Date:

January 2024

Notes:

Pretrained language models released; instruction-finetuned models to follow.

LLM Name	Normistral 7B Warm
Repository 🤗	https://huggingface.co/norallm/normistral-7b-warm
Model Size	7b
Required VRAM	14.5 GB
Updated	2026-08-01
Maintainer	norallm
Model Type	mistral
Model Files	4.8 GB: 1-of-3 4.8 GB: 2-of-3 4.9 GB: 3-of-3 3.5 GB 4.4 GB 5.1 GB 6.0 GB 7.7 GB 14.5 GB
Supported Languages	no nb nn
GGUF Quantization	Yes
Quantization Type	gguf
Model Architecture	MistralForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.34.0.dev0
Tokenizer Class	PreTrainedTokenizerFast
Vocabulary Size	32768
Torch Data Type	bfloat16

Rank the Normistral 7B Warm 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 Normistral 7B Warm

Best Alternatives	Context / RAM	Downloads	Likes
MegaBeam Mistral 7B 300K Gguf	282K / 5 GB	35	3
MegaBeam Mistral 7B 300K Gguf	282K / 5 GB	34	3
Backrooms Mistral 7B 10e	32K / 14.5 GB	128	0
Mistral 7B Instruct V0.2.gguf	32K / 14.5 GB	17	3
Mistral 7B Instruct V0.2 GGUF	32K / 4.4 GB	26	2
...andle Mistral 7B Instruct V0.2	32K / 14.4 GB	103	0
...andle Dolphin 2.2.1 Mistral 7B	32K / 14.4 GB	80	0
Boptruth Agatha 7B	32K / 14.4 GB	129	0
Mahou 1.2a Mistral 7B	32K / 14.4 GB	33	7
Tamil Mistral 7B Instruct V0.1	32K / 14.8 GB	350	14

Note: green Score (e.g. "73.2") means that the model is better than norallm/normistral-7b-warm.