Views
No views yet
tanziro/bd-legal-qwen35-9b-lora, merged into the Qwen/Qwen3.5-9B base model. The merged adapter source used for this GGUF was repair-v2-citation-final-adapter from the source repository.bd-legal-qwen35-9b-lora.Q4_K_M.gguf - final text model GGUF.mmproj-Qwen3.5-9B-BF16.gguf - multimodal projector copied from lmstudio-community/Qwen3.5-9B-GGUF.q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, and down_proj), so the compatible base mmproj is included with attribution instead of generating a new projector. No merged Transformers checkpoint or temporary conversion artifact is included.--no-mtp; this avoids the invalid extra-block layout that can cause missing-tensor load errors in llama.cpp and LM Studio.lms get https://huggingface.co/tanziro/bd-legal-qwen35-9b-lora-GGUFllama-cli -m bd-legal-qwen35-9b-lora.Q4_K_M.gguf -p "Explain why Bangladesh legal citations must be verified."llama-mtmd-cli -m bd-legal-qwen35-9b-lora.Q4_K_M.gguf --mmproj mmproj-Qwen3.5-9B-BF16.gguf --image image.png -p "Describe the image."general.architecture = qwen35, qwen35.block_count = 32, general.file_type = 15, GGUF.version = 3.llama-cli loaded the Q4_K_M GGUF and generated from a short legal-citation prompt.llama-mtmd-cli loaded the GGUF with mmproj-Qwen3.5-9B-BF16.gguf and correctly identified a red test image as red.