LLM EXPLORER 56,604 MODELS INDEXED

Bumblebee Light by MonolithFoundation

By MonolithFoundation · 5 downloads
HF Repo ↗ Copy URL

Bumblebee Light is an open-source language model by MonolithFoundation. Features: 7b LLM, VRAM: 18.6GB, Context: 32K, LLM Explorer Score: 0.12.

  Conversational   Endpoints compatible   Llava llama   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/MonolithFoundation/Bumblebee-Light 

Bumblebee Light Parameters and Internals

Model Type 
VLM, MLLM, Multi-Modal LLM
Use Cases 
Areas:
Research, Commercial applications
Applications:
Chinese OCR understanding, Reasoning, Table and Graphic comprehension, Math calculation
Additional Notes 
Bumblebee is open-source and its weights are available for research purposes to aid the democratization of multi-modal open-source community.
Supported Languages 
zh (Superior performance)
Training Details 
Data Sources:
HuggingFaceFW/fineweb, YanweiLi/MGM-Instruction
Context Length:
576
Model Architecture:
14B LLM-driven multi-modal model, similar to LLaVA architecture with MGM's dual tower visual information input
Input Output 
Accepted Modalities:
text, visual
Release Notes 
Version:
2024.04.25
Date:
2024-04-25
Notes:
First version of Bumblebee released.
LLM NameBumblebee Light
Repository 🤗https://huggingface.co/MonolithFoundation/Bumblebee-Light 
Model Size7b
Required VRAM18.6 GB
Updated2026-08-03
MaintainerMonolithFoundation
Model Typellava_llama
Model Files  5.0 GB: 1-of-4   5.0 GB: 2-of-4   5.0 GB: 3-of-4   3.6 GB: 4-of-4
Model ArchitectureLlavaQwen2ForCausalLM
Context Length32768
Model Max Length32768
Transformers Version4.39.3
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151646
Torch Data Typefloat16
Errorsreplace

Best Alternatives to Bumblebee Light

Best Alternatives
Context / RAM
Downloads
Likes
HuatuoGPT Vision 7B128K / 15.9 GB163230
FastVLM 7B32K / 15.5 GB1124270
Note: green Score (e.g. "73.2") means that the model is better than MonolithFoundation/Bumblebee-Light.