Muse-Glimmer-30B · file← Muse-Glimmer-30B
vision projector
Meta's vision add-on, which lets the model read images. Deleted in August. Mmproj-kquant, Meta.
- Size
- not recorded
- Status
- Deleted 26 Aug 2026
- Scored runs
- 0
ResultsNever scored
This file never completed a scored run. Its record below says why.
The recordHow it was made · how it was served · what went wrong · what happened to it
Rebuilt from the lab's records by Claude, an AI model. The results above come straight from the automated runs. How these pages are made.
| File | mmproj-kquant.gguf |
| Format | GGUF projector, K-quant |
| Where from | downloaded ~2026-08-11 |
| Status | deleted 2026-08-26 |
| Hugging Face | meta-models/Muse-Glimmer-30B-GGUF |
Where it came from
meta-models/Muse-Glimmer-30B-GGUF,mmproj-kquant.gguf, in the Augustget.shrun. Size: not on record (the September repo listing shows a 1.3 GB K-quant projector under bothmmproj-kquant.ggufon unsloth andmmproj-Muse-Glimmer-30B-Q4_K_M.ggufon Meta). Date and revision: not on record. Weights hash: the August files were verified against HF LFS OIDs (lab notes) and the digests were not kept.
How it was served
| label | dates (UTC) | binary and version | ctx / -np / KV / -fa / --jinja / --reasoning-format | harness profile | why this label exists |
|---|---|---|---|---|---|
| none | August 2026 | llama.cpp b10369 | mmproj = …/mmproj-kquant.gguf in preset [muse-glimmer-30b] | none | vision path for the preset |
Seats held: the router preset only (see Muse-Glimmer 30B kquant-dynamic Q4_K (Meta GGUF)).
Decisions
Kept beside the weights; deleted with them on 2026-08-26.
What went wrong
- The mtmd
--jinjahard-crash (this custom template is not supported, try using --jinja) was found with this projector (a lab note);llama-serverenables jinja by default, llama-mtmd-cli does not. - Standing checks: no run-log entries.
What happened to it
Downloaded ~2026-08-11 → preset → deleted 2026-08-26. No RESTORE entry. Re-downloadable from the same repo.