What are the hardware requirements for Table Llava V1.5 13B?

Table Llava V1.5 13B requires approximately 26.6 GB of VRAM and supports a context window of 4K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Table Llava V1.5 13B and how large is it?

Table Llava V1.5 13B is developed by SpursgoZmy, a model with 13b 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 Table Llava V1.5 13B?

Table Llava V1.5 13B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Table Llava V1.5 13B by SpursgoZmy — VRAM 26.6GB, 4K context

Name: Table Llava V1.5 13B
Author: SpursgoZmy

Table Llava V1.5 13B is an open-source language model by SpursgoZmy. Features: 13b LLM, VRAM: 26.6GB, Context: 4K, Instruction-Based, LLM Explorer Score: 0.13.

Arxiv:2310.03744 Arxiv:2406.08100 Dataset:liuhaotian/llava-instr... Dataset:liuhaotian/llava-pretr... Dataset:spursgozmy/mmtab En Endpoints compatible Instruct Llava Pytorch Region:us Sharded

Model Card on HF 🤗: https://huggingface.co/SpursgoZmy/table-llava-v1.5-13b

Table Llava V1.5 13B Parameters and Internals

Model Type

multimodal, chatbot

Use Cases

Areas:

research on large multimodal models and chatbots

Applications:

multimodal table understanding

Primary Use Cases:

research in computer vision, NLP, ML, AI

Limitations:

Limited to one table image as input, low input resolution may limit capacity.

Additional Notes

Evaluated on 17 held-in and 7 held-out tabular benchmarks, and 2 non-tabular benchmarks: TextVQA and llava-bench-in-the-wild.

Supported Languages

en (high)

Training Details

Data Sources:

SpursgoZmy/MMTab, liuhaotian/LLaVA-Instruct-150K, liuhaotian/LLaVA-Pretrain

Data Volume:

Approximately 1.5 million instances

Methodology:

Two-stage training: pre-training with image-caption and table recognition data, followed by instruction tuning with multimodal data.

Model Architecture:

Follows LLaVA-v1.5 with CLIP-ViT-L-336px as the visual encoder, Vicuna-v1.5-13B as base LLM, and a two-layer MLP as the vision-language connector.

Input Output

Input Format:

Single table image of resolution 336*336

Accepted Modalities:

image

LLM Name	Table Llava V1.5 13B
Repository 🤗	https://huggingface.co/SpursgoZmy/table-llava-v1.5-13b
Model Size	13b
Required VRAM	26.6 GB
Updated	2026-07-28
Maintainer	SpursgoZmy
Model Type	llava
Instruction-Based	Yes
Model Files	9.9 GB: 1-of-3 9.9 GB: 2-of-3 6.8 GB: 3-of-3
Supported Languages	en
Model Architecture	LlavaLlamaForCausalLM
Context Length	4096
Model Max Length	4096
Transformers Version	4.31.0
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the Table Llava V1.5 13B 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 Table Llava V1.5 13B

Best Alternatives	Context / RAM	Downloads	Likes
...t4V LLaVA Instruct Mix880k 13B	4K / 26.6 GB	8	2