From 6a22c2dfdaad1595eb098385e225d2cb9f5a58ba Mon Sep 17 00:00:00 2001 From: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com> Date: Tue, 15 Sep 2026 00:06:26 +0000 Subject: [PATCH] fix(gallery): repair and extend Qwopus Flash builds Replace the unavailable Hugging Face revision and update the hashes. Add Q5 and Q6 variants alongside Q4 and Q8, with installation docs. Assisted-by: Codex:GPT-6 --- docs/content/features/model-gallery.md | 18 ++++ gallery/index.yaml | 120 ++++++++++++++++++++++--- 2 files changed, 126 insertions(+), 12 deletions(-) diff --git a/docs/content/features/model-gallery.md b/docs/content/features/model-gallery.md index f0088f583..3d5cd69d2 100644 --- a/docs/content/features/model-gallery.md +++ b/docs/content/features/model-gallery.md @@ -43,6 +43,24 @@ Both views use the same model selection and store the view, search, filter, and selection in the URL. Installing from Explore does not move you away from the catalog; the entry updates in place when the operation finishes. +## Qwopus3.8-27B-Flash + +Install [Qwopus3.8-27B-Flash](https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF) +with automatic selection between Q4_K_M, Q5_K_M, Q6_K, and Q8_0 GGUF builds: + +```bash +local-ai models install qwopus3.8-27b-flash +``` + +To select Q5_K_M explicitly: + +```bash +local-ai models install qwopus3.8-27b-flash --variant qwopus3.8-27b-flash-q5 +``` + +Each build includes the F32 vision projector and uses the embedded chat +template with MTP speculative decoding. Use a current llama.cpp backend. + ## Spark-X2.5-1.7B Install Spark-X2.5-1.7B with automatic selection between its Q4_K_M and Q8_0 diff --git a/gallery/index.yaml b/gallery/index.yaml index 104827cc3..4963e2435 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -156,6 +156,8 @@ uri: https://huggingface.co/unsloth/DeepSeek-V4-Flash-Vision-Exp-GGUF - name: "qwopus3.8-27b-flash" variants: + - model: qwopus3.8-27b-flash-q5 + - model: qwopus3.8-27b-flash-q6 - model: qwopus3.8-27b-flash-q8 url: "github:mudler/LocalAI/gallery/virtual.yaml@master" urls: @@ -186,7 +188,7 @@ disable: true known_usecases: - chat - mmproj: llama-cpp/mmproj/Qwopus3.8-27B-Flash-MTP-Q4_K_M/mmproj-F32.gguf + mmproj: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf options: - use_jinja:true - spec_type:draft-mtp @@ -198,11 +200,11 @@ use_tokenizer_template: true files: - filename: llama-cpp/models/Qwopus3.8-27B-Flash-MTP-Q4_K_M/Qwopus3.8-27B-Flash-MTP-Q4_K_M.gguf - sha256: 2b9a335bd33bb977d3127062061c12cc94f37ceef1850d372317fbf4ee89d9c5 - uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/e146d61e88782677805b3b68ad3adf8674dde80d/Qwopus3.8-27B-Flash-MTP-Q4_K_M.gguf - - filename: llama-cpp/mmproj/Qwopus3.8-27B-Flash-MTP-Q4_K_M/mmproj-F32.gguf - sha256: be3f444ecaedf084fd19399ba18cefcbbe3789bb81f4ad5ed67c91a1a3f48a67 - uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/e146d61e88782677805b3b68ad3adf8674dde80d/mmproj-F32.gguf + sha256: 1b76ad1596d6afc0b6e38a2b45b404462920739bd3087960b7a18a74ca938a74 + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/Qwopus3.8-27B-Flash-MTP-Q4_K_M.gguf + - filename: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf + sha256: 52e6818e4d18eea010c50e5245eaa10a8cc3dcc30efea4ff60cbad8abf5669e1 + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/mmproj-F32.gguf - name: "qwopus3.8-27b-flash-q8" url: "github:mudler/LocalAI/gallery/virtual.yaml@master" urls: @@ -233,7 +235,7 @@ disable: true known_usecases: - chat - mmproj: llama-cpp/mmproj/Qwopus3.8-27B-Flash-MTP-Q4_K_M/mmproj-F32.gguf + mmproj: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf options: - use_jinja:true - spec_type:draft-mtp @@ -245,11 +247,105 @@ use_tokenizer_template: true files: - filename: llama-cpp/models/Qwopus3.8-27B-Flash-MTP-Q8_0/Qwopus3.8-27B-Flash-MTP-Q8_0.gguf - sha256: a97de033ea60066a1941fe4dfb2e631b390cbc7c7563cea2446b7f08065228ec - uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/e146d61e88782677805b3b68ad3adf8674dde80d/Qwopus3.8-27B-Flash-MTP-Q8_0.gguf - - filename: llama-cpp/mmproj/Qwopus3.8-27B-Flash-MTP-Q4_K_M/mmproj-F32.gguf - sha256: be3f444ecaedf084fd19399ba18cefcbbe3789bb81f4ad5ed67c91a1a3f48a67 - uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/e146d61e88782677805b3b68ad3adf8674dde80d/mmproj-F32.gguf + sha256: 98e47b3da5f41c1753a45dbbc4d98e94fd05101c1ab5425d341ff0e58fe2e292 + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/Qwopus3.8-27B-Flash-MTP-Q8_0.gguf + - filename: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf + sha256: 52e6818e4d18eea010c50e5245eaa10a8cc3dcc30efea4ff60cbad8abf5669e1 + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/mmproj-F32.gguf +- name: "qwopus3.8-27b-flash-q5" + url: "github:mudler/LocalAI/gallery/virtual.yaml@master" + urls: + - https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash + - https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF + description: | + Qwopus3.8-27B-Flash is a Qwen3.8-27B fine-tune for reasoning and agent + workloads. This Q5_K_M GGUF includes the F32 vision projector and uses + llama.cpp's embedded chat template with MTP speculative decoding. + The publisher reports a known Python code indentation issue. + license: "apache-2.0" + tags: + - llm + - gguf + - qwen + - qwen3 + - vision + - multimodal + - instruction-tuned + - reasoning + - mtp + icon: https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/CI_Demo/mathv-1327.jpg + overrides: + backend: llama-cpp + function: + automatic_tool_parsing_fallback: true + grammar: + disable: true + known_usecases: + - chat + mmproj: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf + options: + - use_jinja:true + - spec_type:draft-mtp + - spec_n_max:6 + - spec_p_min:0.75 + parameters: + model: llama-cpp/models/Qwopus3.8-27B-Flash-MTP-Q5_K_M/Qwopus3.8-27B-Flash-MTP-Q5_K_M.gguf + template: + use_tokenizer_template: true + files: + - filename: llama-cpp/models/Qwopus3.8-27B-Flash-MTP-Q5_K_M/Qwopus3.8-27B-Flash-MTP-Q5_K_M.gguf + sha256: abc97cf74cae75be61667f3c640d9abc76ad87c617a10ee1fa47beb66aec3746 + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/Qwopus3.8-27B-Flash-MTP-Q5_K_M.gguf + - filename: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf + sha256: 52e6818e4d18eea010c50e5245eaa10a8cc3dcc30efea4ff60cbad8abf5669e1 + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/mmproj-F32.gguf +- name: "qwopus3.8-27b-flash-q6" + url: "github:mudler/LocalAI/gallery/virtual.yaml@master" + urls: + - https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash + - https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF + description: | + Qwopus3.8-27B-Flash is a Qwen3.8-27B fine-tune for reasoning and agent + workloads. This Q6_K GGUF includes the F32 vision projector and uses + llama.cpp's embedded chat template with MTP speculative decoding. + The publisher reports a known Python code indentation issue. + license: "apache-2.0" + tags: + - llm + - gguf + - qwen + - qwen3 + - vision + - multimodal + - instruction-tuned + - reasoning + - mtp + icon: https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/CI_Demo/mathv-1327.jpg + overrides: + backend: llama-cpp + function: + automatic_tool_parsing_fallback: true + grammar: + disable: true + known_usecases: + - chat + mmproj: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf + options: + - use_jinja:true + - spec_type:draft-mtp + - spec_n_max:6 + - spec_p_min:0.75 + parameters: + model: llama-cpp/models/Qwopus3.8-27B-Flash-MTP-Q6_K/Qwopus3.8-27B-Flash-MTP-Q6_K.gguf + template: + use_tokenizer_template: true + files: + - filename: llama-cpp/models/Qwopus3.8-27B-Flash-MTP-Q6_K/Qwopus3.8-27B-Flash-MTP-Q6_K.gguf + sha256: e41fb287a3669d92a131c31629582ef3e9e1df61e9d735fe6b1c71cb9994f7ca + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/Qwopus3.8-27B-Flash-MTP-Q6_K.gguf + - filename: llama-cpp/mmproj/qwopus3.8-27b-flash-9b90cd22/mmproj-F32.gguf + sha256: 52e6818e4d18eea010c50e5245eaa10a8cc3dcc30efea4ff60cbad8abf5669e1 + uri: https://huggingface.co/Jackrong/Qwopus3.8-27B-Flash-GGUF/resolve/9b90cd22b19276bc9e75c0e6d1665fa48619dd5b/mmproj-F32.gguf - name: "qwen3.8-27b" url: "github:mudler/LocalAI/gallery/virtual.yaml@master" urls: