---
license: apache-2.0
base_model: openbmb/MiniCPM5-2B-GGUF
library_name: gguf
pipeline_tag: text-generation
tags:
- webgpu
- wllama
- minicpm5
- gguf
- q8_0
---
# MiniCPM5-2B · Q8_0 browser shards

A **lossless GGUF split** of OpenBMB's official [MiniCPM5-2B-Q8_0.gguf](https://huggingface.co/openbmb/MiniCPM5-2B-GGUF).
No re-quantization, pruning, conversion of tensor values, or changes to the model architecture.
All 381 tensors were compared by SHA-256 against the original. See `verification.json`.

Split with `llama-gguf-split --split-max-size 512M` into six files, each under the browser's 2 GB ArrayBuffer limit.
Pass the first shard URL to Wllama 3.6.1; it discovers and downloads the other five in parallel.

**Chat and create files entirely in your browser:** [MiniCPM5 WebGPU](https://huggingface.co/spaces/ProCreations/minicpm5-2b-webgpu)

Model by [OpenBMB](https://huggingface.co/openbmb). Browser experience by [ProCreations](https://huggingface.co/ProCreations).
Original weights are Apache 2.0. Please refer to the upstream model card for model capabilities, training, and limitations.