Skip to content

Commit d28ccf3

Browse files
gallery: add Qwen3.6 14B FableVibes variants (#11317)
Add Q4_K_M and Q8_0 llama.cpp entries with the shared Q8_0 multimodal projector. Assisted-by: Codex:gpt-5 Co-authored-by: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com>
1 parent 95bd59d commit d28ccf3

1 file changed

Lines changed: 79 additions & 0 deletions

File tree

gallery/index.yaml

Lines changed: 79 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1387,6 +1387,85 @@
13871387
- filename: llama-cpp/mmproj/Qwythos-9B-v2-MTP-Q4_K_M/mmproj-Qwythos-9B-v2-BF16.gguf
13881388
sha256: 0d1687cb33124c78acab788b342d4a2eaf85b3035e87c3abe4ee9d0b84ddb4f5
13891389
uri: https://huggingface.co/empero-ai/Qwythos-9B-v2-GGUF/resolve/main/mmproj-Qwythos-9B-v2-BF16.gguf
1390+
- &qwen3-6-14b-a3b-fablevibes
1391+
name: "qwen3.6-14b-a3b-fablevibes"
1392+
variants:
1393+
- model: qwen3.6-14b-a3b-fablevibes-q8
1394+
url: "github:mudler/LocalAI/gallery/virtual.yaml@master"
1395+
urls:
1396+
- https://huggingface.co/tvall43/Qwen3.6-14B-A3B-FableVibes
1397+
- https://huggingface.co/tvall43/Qwen3.6-14B-A3B-FableVibes-GGUF
1398+
description: |
1399+
Qwen3.6-14B-A3B-FableVibes is an Apache-2.0 mixture-of-experts reasoning
1400+
model distilled from Fable 5 and Claude Opus traces, with additional tool
1401+
calling and coding data. It retains Qwen 3.6 vision support while pruning
1402+
the 35B-A3B base to a 14B consumer-oriented footprint. This default entry
1403+
uses the recommended Q4_K_M GGUF quantization and its Q8_0 multimodal
1404+
projector.
1405+
license: "apache-2.0"
1406+
tags:
1407+
- llm
1408+
- gguf
1409+
- cpu
1410+
- gpu
1411+
- moe
1412+
- reasoning
1413+
- thinking
1414+
- vision
1415+
- multimodal
1416+
last_checked: "2026-08-03"
1417+
overrides:
1418+
backend: llama-cpp
1419+
function:
1420+
automatic_tool_parsing_fallback: true
1421+
grammar:
1422+
disable: true
1423+
known_usecases:
1424+
- chat
1425+
mmproj: llama-cpp/mmproj/Qwen3.6-14B-A3B-FableVibes/Qwen3.6-14B-A3B-FableVibes-mmproj-Q8_0.gguf
1426+
options:
1427+
- use_jinja:true
1428+
parameters:
1429+
model: llama-cpp/models/Qwen3.6-14B-A3B-FableVibes/Qwen3.6-14B-A3B-FableVibes-Q4_K_M.gguf
1430+
template:
1431+
use_tokenizer_template: true
1432+
files:
1433+
- filename: llama-cpp/models/Qwen3.6-14B-A3B-FableVibes/Qwen3.6-14B-A3B-FableVibes-Q4_K_M.gguf
1434+
sha256: 21aa4b0b28090469e8a319c889451df2f1ea6aad27ac3818c8c8a86f86d5bc9e
1435+
uri: huggingface://tvall43/Qwen3.6-14B-A3B-FableVibes-GGUF/Qwen3.6-14B-A3B-FableVibes-Q4_K_M.gguf
1436+
- filename: llama-cpp/mmproj/Qwen3.6-14B-A3B-FableVibes/Qwen3.6-14B-A3B-FableVibes-mmproj-Q8_0.gguf
1437+
sha256: ca27dbf0c65a7232e9458bfdda8bc45efc09ab60e4cc6f58ea0c7b7cc2253257
1438+
uri: huggingface://tvall43/Qwen3.6-14B-A3B-FableVibes-GGUF/Qwen3.6-14B-A3B-FableVibes-mmproj-Q8_0.gguf
1439+
- !!merge <<: *qwen3-6-14b-a3b-fablevibes
1440+
name: "qwen3.6-14b-a3b-fablevibes-q8"
1441+
variants: []
1442+
description: |
1443+
Qwen3.6-14B-A3B-FableVibes is an Apache-2.0 mixture-of-experts reasoning
1444+
model distilled from Fable 5 and Claude Opus traces, with additional tool
1445+
calling and coding data. This entry uses the near-lossless Q8_0 GGUF
1446+
quantization and its matching Q8_0 multimodal projector.
1447+
overrides:
1448+
backend: llama-cpp
1449+
function:
1450+
automatic_tool_parsing_fallback: true
1451+
grammar:
1452+
disable: true
1453+
known_usecases:
1454+
- chat
1455+
mmproj: llama-cpp/mmproj/Qwen3.6-14B-A3B-FableVibes-Q8_0/Qwen3.6-14B-A3B-FableVibes-mmproj-Q8_0.gguf
1456+
options:
1457+
- use_jinja:true
1458+
parameters:
1459+
model: llama-cpp/models/Qwen3.6-14B-A3B-FableVibes-Q8_0/Qwen3.6-14B-A3B-FableVibes-Q8_0.gguf
1460+
template:
1461+
use_tokenizer_template: true
1462+
files:
1463+
- filename: llama-cpp/models/Qwen3.6-14B-A3B-FableVibes-Q8_0/Qwen3.6-14B-A3B-FableVibes-Q8_0.gguf
1464+
sha256: ddea86093b863215fa75969d81df494fafdb0e6c4a65af557ed3a710e2238e58
1465+
uri: huggingface://tvall43/Qwen3.6-14B-A3B-FableVibes-GGUF/Qwen3.6-14B-A3B-FableVibes-Q8_0.gguf
1466+
- filename: llama-cpp/mmproj/Qwen3.6-14B-A3B-FableVibes-Q8_0/Qwen3.6-14B-A3B-FableVibes-mmproj-Q8_0.gguf
1467+
sha256: ca27dbf0c65a7232e9458bfdda8bc45efc09ab60e4cc6f58ea0c7b7cc2253257
1468+
uri: huggingface://tvall43/Qwen3.6-14B-A3B-FableVibes-GGUF/Qwen3.6-14B-A3B-FableVibes-mmproj-Q8_0.gguf
13901469
- name: "qwen3.6-27b-fable-fusion-711-uncensored-heretic-nm-dau-neo-max-mtp"
13911470
url: "github:mudler/LocalAI/gallery/virtual.yaml@master"
13921471
urls:

0 commit comments

Comments
 (0)