HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

Qwen2.5-VL-72B-Instruct-bnb-4bit

unsloth/Qwen2.5-VL-72B-Instruct-bnb-4bit

kind
bnb-4bit
quant
4bit
family
qwen
modality
image-text-to-text
root model
Qwen/Qwen2.5-VL-72B-Instruct
provider
unsloth
license
other
language
en
context
32,768
library
transformers
vRAM (GB)
—
size on disk
—
downloads
266
likes
15
created
—
last modified
2025-05-12T09:33:41+00:00 (2025)
huggingface
https://huggingface.co/unsloth/Qwen2.5-VL-72B-Instruct-bnb-4bit
tags
transformerssafetensorsqwen2_5_vlimage-text-to-textmultimodalunslothconversationalenarxiv:2309.00071arxiv:2409.12191arxiv:2308.12966base_model:Qwen/Qwen2.5-VL-72B-Instructbase_model:quantized:Qwen/Qwen2.5-VL-72B-Instructlicense:othertext-generation-inferenceendpoints_compatible4-bitbitsandbytesdeploy:azureregion:us

Benchmark scores (7)

benchmarkcategoryscorecontextconfevidence
DocVQAmultimodal91.1DocVQA_VALmedium| DocVQA_VAL | 91.1 | 95.2 | 92.1 | 96.5 | 96.1 | 96.4 |
ChartQAmultimodal86.7ChartQA_TESTmedium| ChartQA_TEST | 86.7 | 90.8 | 85.2 | 88.3 | 88.3 | 89.5 |
MMBenchmultimodal82.1MMBench_DEV_EN_V11medium| MMBench_DEV_EN_V11 | 82.1 | 83.4 | 83.0 | 88.5 | 86.6 | 88 |
MMMUmultimodal70.3MMMU<sub>val</sub>medium| MMMU<sub>val</sub> | 70.3 | 70.4 | 70.7 | 70.1 | 64.5 | 70.2 |
MathVistamultimodal63.8MathVista_MINImedium| MathVista_MINI | 63.8 | 65.4 | 73.1 | 76.6 | 70.5 | 74.8 |
OSWorldagentic14.9OSWorldmedium| OSWorld | | | 14.90 | 10.26 | | 8.83 |
MMBenchmultimodal1.63MMBench-Videomedium| MMBench-Video | 1.63 | 1.30 | 1.97 | 1.70 | 2.02 |

Discussions (0)

none scraped.

Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.