You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

D1AVG bf16 (.pqm)

D1AVG_bf16.pqm -- 35B MoE, .pqm container, bf16 weights. 69367107584 bytes, published in 22 parts (HF's per-file limit is 50 GB).

Reassemble

cat D1AVG_bf16.pqm.part-* > D1AVG_bf16.pqm
md5sum -c SHASUMS.txt

Measured quality (MalayMMLU 0-shot, 500-item stratified sample, generate+extract)

accuracy 0.826, null 0.002 -- Humanities .86 / Language .89 / Others .77 / STEM .84 / SocSci .77

Serving

Needs the .pqm-aware engine with PRISMX_PQM_STANDALONE=1; tokenizer is D1AVG.tok in this repo. On a 32 GB card it serves with an expert host-pinning plan (--gpu-mem auto).

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support