LLM EXPLORER 55,709 MODELS INDEXED

Llama 3.2 90B Vision Instruct 4bit by mlx-community

By mlx-community · 318 downloads
HF Repo ↗ Copy URL

Llama 3.2 90B Vision Instruct 4bit is an open-source language model by mlx-community. Features: 90b LLM, VRAM: 49.9GB, License: llama3.2, Quantized, Instruction-Based.

  4bit Base model:finetune:meta-llama... Base model:meta-llama/llama-3....   Conversational   Deploy:azure   En   Endpoints compatible   Facebook   Image-text-to-text   Instruct   Llama   Llama-3   Meta   Mllama   Mlx   Multimodal   Quantized   Region:us   Safetensors   Sharded   Tensorflow   Unsloth   Vision
Model Card on HF 🤗: https://huggingface.co/mlx-community/Llama-3.2-90B-Vision-Instruct-4bit 

Llama 3.2 90B Vision Instruct 4bit Parameters and Internals

LLM NameLlama 3.2 90B Vision Instruct 4bit
Repository 🤗https://huggingface.co/mlx-community/Llama-3.2-90B-Vision-Instruct-4bit 
Base Model(s)  meta-llama/Llama-3.2-90B-Vision-Instruct   meta-llama/Llama-3.2-90B-Vision-Instruct
Model Size90b
Required VRAM49.9 GB
Updated2026-08-08
Maintainermlx-community
Model Typemllama
Instruction-BasedYes
Model Files  5.3 GB: 1-of-10   5.3 GB: 2-of-10   5.3 GB: 3-of-10   5.3 GB: 4-of-10   5.3 GB: 5-of-10   5.3 GB: 6-of-10   5.3 GB: 7-of-10   5.3 GB: 8-of-10   5.3 GB: 9-of-10   2.2 GB: 10-of-10
Supported Languagesen
Quantization Type4bit
Model ArchitectureMllamaForConditionalGeneration
Licensellama3.2
Model Max Length131072
Transformers Version4.45.0
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<|finetune_right_pad_id|>
Torch Data Typebfloat16