What are the hardware requirements for OpenMistral MoE?

OpenMistral MoE requires approximately 48.3 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed OpenMistral MoE and how large is it?

OpenMistral MoE is developed by Yash21, a model with 7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does OpenMistral MoE perform on standard benchmarks?

OpenMistral MoE has the following published scores: MMLU 60.69. Compare against reference models on this page or on the LLM Explorer leaderboards.

OpenMistral MoE by Yash21 — VRAM 48.3GB, 32K context

Name: OpenMistral MoE
Author: Yash21

OpenMistral MoE is an open-source language model by Yash21. Features: 7b LLM, VRAM: 48.3GB, Context: 32K, Instruction-Based, LLM Explorer Score: 0.13, Arc: 64.1, HellaSwag: 84, MMLU: 60.7, GSM8K: 58.4.

Autotrain compatible Conversational Endpoints compatible Instruct Mixtral Moe Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Yash21/OpenMistral-MoE

OpenMistral MoE Parameters and Internals

Model Type

text-generation

Input Output

Accepted Modalities:

text

LLM Name	OpenMistral MoE
Repository 🤗	https://huggingface.co/Yash21/OpenMistral-MoE
Model Size	7b
Required VRAM	48.3 GB
Updated	2024-10-29
Maintainer	Yash21
Model Type	mixtral
Instruction-Based	Yes
Model Files	9.9 GB: 1-of-5 10.0 GB: 2-of-5 10.0 GB: 3-of-5 10.0 GB: 4-of-5 8.4 GB: 5-of-5
Model Architecture	MixtralForCausalLM
Context Length	32768
Model Max Length	32768
Transformers Version	4.36.2
Tokenizer Class	LlamaTokenizer
Padding Token	<s>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the OpenMistral MoE 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 OpenMistral MoE

Best Alternatives	Context / RAM	Downloads	Likes
Mini Mixtral V0.2	32K / 25.8 GB	3923	4
Multilingual Mistral	32K / 93.5 GB	65	2
Magiq 3	32K / 37.1 GB	5	3
FNCARL9000	32K / 48.3 GB	8	0
Bigstral 12B 32K 8xMoE	32K / 163.3 GB	12	2
CollAIborate4x7B	32K / 48.7 GB	6	1
CollAIborate4x7B	32K / 48.7 GB	1	1
OpenMistral MoE	32K / 48.3 GB	6	0
...t V0.2 2x7B MoE 6.0bpw H6 EXL2	32K / 9.9 GB	3	1

Note: green Score (e.g. "73.2") means that the model is better than Yash21/OpenMistral-MoE.