What are the hardware requirements for Llama Joycaption Beta One Hf Llava?

Llama Joycaption Beta One Hf Llava requires approximately 17 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Llama Joycaption Beta One Hf Llava and how large is it?

Llama Joycaption Beta One Hf Llava is developed by fancyfeast, a model with 8b 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 Llama Joycaption Beta One Hf Llava?

Llama Joycaption Beta One Hf Llava is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Llama Joycaption Beta One Hf Llava by fancyfeast — VRAM 17GB

Name: Llama Joycaption Beta One Hf Llava
Author: fancyfeast

Llama Joycaption Beta One Hf Llava is an open-source language model by fancyfeast. Features: 8b LLM, VRAM: 17GB, LLM Explorer Score: 0.28.

Base model:finetune:google/sig... Base model:google/siglip2-so40... Captioning Conversational Endpoints compatible Image-text-to-text Llava Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/fancyfeast/llama-joycaption-beta-one-hf-llava

Llama Joycaption Beta One Hf Llava Parameters and Internals

LLM Name	Llama Joycaption Beta One Hf Llava
Repository 🤗	https://huggingface.co/fancyfeast/llama-joycaption-beta-one-hf-llava
Base Model(s)	meta-llama/Llama-3.1-8B-Instruct Siglip2 So400m Patch14 384 meta-llama/Llama-3.1-8B-Instruct google/siglip2-so400m-patch14-384
Model Size	8b
Required VRAM	17 GB
Updated	2026-08-13
Maintainer	fancyfeast
Model Type	llava
Model Files	4.9 GB: 1-of-4 5.0 GB: 2-of-4 4.9 GB: 3-of-4 2.2 GB: 4-of-4
Model Architecture	LlavaForConditionalGeneration
Model Max Length	131072
Transformers Version	4.51.0.dev0
Tokenizer Class	PreTrainedTokenizer
Torch Data Type	bfloat16

Rank the Llama Joycaption Beta One Hf Llava 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 Llama Joycaption Beta One Hf Llava

Best Alternatives	Context / RAM	Downloads	Likes
Llava Llama 3 8B V1.1 4bit	0K / 5.5 GB	58	4
Llava Llama 3 8B V1.1 8bit	0K / 9.4 GB	22	4