mirror of
https://github.com/mudler/LocalAI.git
synced 2026-10-28 09:08:54 -04:00
chore(gallery): add Turbo Brilliance 50X variants
Add Q4_K_M and Q8_0 builds with the F16 vision projector for llama.cpp. Pin downloads and verify SHA256 against HF metadata and response headers. Document variant selection and keep the embedded chat template. Assisted-by: Codex:GPT-6
This commit is contained in:
localai-org-maint-bot committed
1 parent
33459f2e2a
commit
a3a910f833
2 files changed
+93
No files matched your search
@@ -187,6 +187,27 @@ To select Q8 explicitly, run `local-ai models install cyber-tiel-coder-35b-a3b-q
|
||||
Both configurations use the embedded chat template and default to 32,768 context tokens.
|
||||
The [model card](https://huggingface.co/peculiar-ragdoll/Cyber-Tiel-Coder-35B-A3B-GGUF-MTP) describes its abliterated Ornith-1.5 base and MIT license.
|
||||
|
||||
## Qwen3.8-27B Turbo Brilliance 50X
|
||||
|
||||
The gallery provides Q4_K_M and Q8_0 builds for llama.cpp. Both include the
|
||||
F16 vision projector and use the model's embedded chat template. The default
|
||||
context is 32,768 tokens. These entries do not enable MTP speculative decoding.
|
||||
|
||||
Install with automatic variant selection:
|
||||
|
||||
```bash
|
||||
local-ai models install qwen3.8-27b-turbo-brilliance-50x-q4
|
||||
```
|
||||
|
||||
To select Q8_0 explicitly:
|
||||
|
||||
```bash
|
||||
local-ai models install --variant qwen3.8-27b-turbo-brilliance-50x-q8 qwen3.8-27b-turbo-brilliance-50x-q4
|
||||
```
|
||||
|
||||
See the [publisher's model card](https://huggingface.co/DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF)
|
||||
for reasoning modes, prompt syntax, and sampling guidance.
|
||||
|
||||
## Qwen3.8-27B Agention Precision
|
||||
|
||||
The gallery includes Agention Precision IQ4_XS and Q4_K_M GGUF builds of
|
||||
|
||||
@@ -843,6 +843,78 @@
|
||||
- filename: Hemmingway-1.Q8_0.gguf
|
||||
sha256: 10b566343db6905215c2d635285ffb43ada05a2f9b1e8e85562c8e366fac1809
|
||||
uri: huggingface://mradermacher/Hemmingway-1-GGUF/Hemmingway-1.Q8_0.gguf
|
||||
- name: qwen3.8-27b-turbo-brilliance-50x-q4
|
||||
variants:
|
||||
- model: qwen3.8-27b-turbo-brilliance-50x-q8
|
||||
url: github:mudler/LocalAI/gallery/qwen3.yaml@master
|
||||
urls:
|
||||
- https://huggingface.co/Qwen/Qwen3.8-27B
|
||||
- https://huggingface.co/DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF
|
||||
description: |
|
||||
DavidAU's Turbo Brilliance 50X edition of Qwen3.8-27B includes a custom
|
||||
chat template with selectable reasoning and instruction modes. This
|
||||
Q4_K_M GGUF build includes the F16 vision projector for image input.
|
||||
Uses the embedded template with a 32,768-token context by default.
|
||||
license: apache-2.0
|
||||
tags:
|
||||
- llm
|
||||
- gguf
|
||||
- cpu
|
||||
- gpu
|
||||
- reasoning
|
||||
- vision
|
||||
overrides:
|
||||
context_size: 32768
|
||||
mmproj: qwen3.8-27b-turbo-brilliance-50x/mmproj-F16.gguf
|
||||
parameters:
|
||||
model: Qwen3.8-27B-Turbo-Brilliance-50X-RI-NEO-Q4_K_M.gguf
|
||||
temperature: 0.7
|
||||
top_k: 64
|
||||
top_p: 0.95
|
||||
min_p: 0.05
|
||||
repeat_penalty: 1
|
||||
files:
|
||||
- filename: Qwen3.8-27B-Turbo-Brilliance-50X-RI-NEO-Q4_K_M.gguf
|
||||
sha256: 2e4fbc3b9c9cd45569dc506a87571a3dedfe213f9138cf4eece8d382c8d3794e
|
||||
uri: https://huggingface.co/DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF/resolve/35d2d177d8efe3acc102afb1fb96b0bc1fb0b5d9/Qwen3.8-27B-Turbo-Brilliance-50X-RI-NEO-Q4_K_M.gguf
|
||||
- filename: qwen3.8-27b-turbo-brilliance-50x/mmproj-F16.gguf
|
||||
sha256: cb0a71fe9b84af2552fc6260acf8e9773a29d115c683aba01c72a7ac537b212b
|
||||
uri: https://huggingface.co/DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF/resolve/35d2d177d8efe3acc102afb1fb96b0bc1fb0b5d9/mmproj-F16.gguf
|
||||
- name: qwen3.8-27b-turbo-brilliance-50x-q8
|
||||
url: github:mudler/LocalAI/gallery/qwen3.yaml@master
|
||||
urls:
|
||||
- https://huggingface.co/Qwen/Qwen3.8-27B
|
||||
- https://huggingface.co/DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF
|
||||
description: |
|
||||
DavidAU's Turbo Brilliance 50X edition of Qwen3.8-27B includes a custom
|
||||
chat template with selectable reasoning and instruction modes. This
|
||||
Q8_0 GGUF build includes the F16 vision projector for image input.
|
||||
Uses the embedded template with a 32,768-token context by default.
|
||||
license: apache-2.0
|
||||
tags:
|
||||
- llm
|
||||
- gguf
|
||||
- cpu
|
||||
- gpu
|
||||
- reasoning
|
||||
- vision
|
||||
overrides:
|
||||
context_size: 32768
|
||||
mmproj: qwen3.8-27b-turbo-brilliance-50x/mmproj-F16.gguf
|
||||
parameters:
|
||||
model: Qwen3.8-27B-Turbo-Brilliance-50X-RI-NEO-Q8_0.gguf
|
||||
temperature: 0.7
|
||||
top_k: 64
|
||||
top_p: 0.95
|
||||
min_p: 0.05
|
||||
repeat_penalty: 1
|
||||
files:
|
||||
- filename: Qwen3.8-27B-Turbo-Brilliance-50X-RI-NEO-Q8_0.gguf
|
||||
sha256: b354f6f683a55bdf17b5c51f0c203dc5c4da38bf877932fe82ff909811728f64
|
||||
uri: https://huggingface.co/DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF/resolve/35d2d177d8efe3acc102afb1fb96b0bc1fb0b5d9/Qwen3.8-27B-Turbo-Brilliance-50X-RI-NEO-Q8_0.gguf
|
||||
- filename: qwen3.8-27b-turbo-brilliance-50x/mmproj-F16.gguf
|
||||
sha256: cb0a71fe9b84af2552fc6260acf8e9773a29d115c683aba01c72a7ac537b212b
|
||||
uri: https://huggingface.co/DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF/resolve/35d2d177d8efe3acc102afb1fb96b0bc1fb0b5d9/mmproj-F16.gguf
|
||||
- name: qwen3.8-27b
|
||||
url: github:mudler/LocalAI/gallery/virtual.yaml@master
|
||||
urls:
|
||||
|
||||
Reference in new issue
Block a user