Quantization recipe
(advanced, embedded in
quantization_config.json
)
Setting
Value
Format
EXL3
Bits per weight
4.0
Head bits
8
Calibration rows
250
Codebook
MCG
Out‑scales
always
Parallel mode
enabled
(MoE expert batching)
Loaded automatically by every ExLlamaV3 loader; reproduced here for searchability.
License & use
Use and license
follow the
base model
.
Quantization adds no additional restrictions. Refer to the upstream repository for terms, citation, and safety documentation.
Quantized with
BlockQuant
· convention
{org}/{model}-exl3-{bpw}bpw
Requirements
First EXL3 of this architecture. Loading needs exllamav3 with OpenPanguV2 support from
Honkware/exllamav3@openpangu
; stock
exllamav3 releases will not load it yet. mHC, MoME conv, attention sink and DSA indexer tensors
ship unquantized.
Multi-token prediction
All three MTP draft layers (46-48) are included and quantized. The per-depth heads and embeddings
are tied to the trunk in the base checkpoint, so they are borrowed at load rather than duplicated.
Teacher-forced acceptance against the quantized trunk: 92.5% / 89.2% / 88.6% by depth (bf16 reference: 89.2% / 88.6% / 85.9%).
Scope
DSA layers run dense in this build. Batch forward and the MTP chain are validated against the
bf16 reference. Cached (paged) generation and per-depth MTP draft dispatch are available on the
pangu-paged branch
: single-stream chat
runs at about 14 tok/s on an H100 and up to 32 tok/s with 3-token MTP drafting (86-96 percent
depth-0 acceptance). Batch size 1 only for now; concurrent-job decoding has a known issue under
investigation.
Quality vs the bf16 original on natural text: mean KL 0.0817, top-1 agreement 92.9%.
License
Use and license follow the base model (LICENSE included in this repo, attribution required).
Powered by openPangu.
Runs of blockblockblock openPangu-2.0-Flash-exl3-4.0bpw on huggingface.co
471
Total runs
0
24-hour runs
0
3-day runs
29
7-day runs
440
30-day runs
More Information About openPangu-2.0-Flash-exl3-4.0bpw huggingface.co Model
More openPangu-2.0-Flash-exl3-4.0bpw license Visit here:
openPangu-2.0-Flash-exl3-4.0bpw huggingface.co is an AI model on huggingface.co that provides openPangu-2.0-Flash-exl3-4.0bpw's model effect (), which can be used instantly with this blockblockblock openPangu-2.0-Flash-exl3-4.0bpw model. huggingface.co supports a free trial of the openPangu-2.0-Flash-exl3-4.0bpw model, and also provides paid use of the openPangu-2.0-Flash-exl3-4.0bpw. Support call openPangu-2.0-Flash-exl3-4.0bpw model through api, including Node.js, Python, http.
openPangu-2.0-Flash-exl3-4.0bpw huggingface.co is an online trial and call api platform, which integrates openPangu-2.0-Flash-exl3-4.0bpw's modeling effects, including api services, and provides a free online trial of openPangu-2.0-Flash-exl3-4.0bpw, you can try openPangu-2.0-Flash-exl3-4.0bpw online for free by clicking the link below.
blockblockblock openPangu-2.0-Flash-exl3-4.0bpw online free url in huggingface.co:
openPangu-2.0-Flash-exl3-4.0bpw is an open source model from GitHub that offers a free installation service, and any user can find openPangu-2.0-Flash-exl3-4.0bpw on GitHub to install. At the same time, huggingface.co provides the effect of openPangu-2.0-Flash-exl3-4.0bpw install, users can directly use openPangu-2.0-Flash-exl3-4.0bpw installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
openPangu-2.0-Flash-exl3-4.0bpw install url in huggingface.co: