What are the hardware requirements for MoE LLaVA Phi2 2.7B 4e?

MoE LLaVA Phi2 2.7B 4e requires approximately 11.2 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 MoE LLaVA Phi2 2.7B 4e and how large is it?

MoE LLaVA Phi2 2.7B 4e is developed by LanguageBind, a model with 5.6b 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 MoE LLaVA Phi2 2.7B 4e?

MoE LLaVA Phi2 2.7B 4e is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

MoE LLaVA Phi2 2.7B 4e by LanguageBind — VRAM 11.2GB, 2K context

Name: MoE LLaVA Phi2 2.7B 4e
Author: LanguageBind

MoE LLaVA Phi2 2.7B 4e is an open-source language model by LanguageBind. Features: 5.6b LLM, VRAM: 11.2GB, Context: 2K, License: apache-2.0, MoE, LLM Explorer Score: 0.11.

Arxiv:2311.10122 Arxiv:2401.15947 Custom code Endpoints compatible Moe Moe llava phi Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/LanguageBind/MoE-LLaVA-Phi2-2.7B-4e

MoE LLaVA Phi2 2.7B 4e Parameters and Internals

Model Type

vision-language, multimodal

Additional Notes

MoE-LLaVA shows excellent performance in multi-modal learning with fewer parameters.

Training Details

Methodology:

A simple MoE tuning stage was added.

Training Time:

2 days

Hardware Used:

8 V100 GPUs

LLM Name	MoE LLaVA Phi2 2.7B 4e
Repository 🤗	https://huggingface.co/LanguageBind/MoE-LLaVA-Phi2-2.7B-4e
Model Size	5.6b
Required VRAM	11.2 GB
Updated	2026-07-22
Maintainer	LanguageBind
Model Type	moe_llava_phi
Model Files	5.0 GB: 1-of-3 5.0 GB: 2-of-3 1.2 GB: 3-of-3
Model Architecture	MoELLaVAPhiForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.36.2
Tokenizer Class	CodeGenTokenizer
Padding Token	<\|extra_0\|>
Vocabulary Size	51200
Torch Data Type	bfloat16
Errors	replace

Rank the MoE LLaVA Phi2 2.7B 4e 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 MoE LLaVA Phi2 2.7B 4e

Best Alternatives	Context / RAM	Downloads	Likes
DynMoE Phi 2 2.7B	2K / 11.2 GB	10	4

Note: green Score (e.g. "73.2") means that the model is better than LanguageBind/MoE-LLaVA-Phi2-2.7B-4e.