Add vision projector (mmproj) for multimodal support

Upload re-quantized IQ4_XS with official chat template embedded

Delete mmproj-Elbaz-GLM-4.6V-Flash-PRISM-F16.gguf

Remove old IQ4_XS (missing chat template in metadata)

Add official GLM-4.6V-Flash chat template

Update tokenizer_config.json with official GLM-4.6V-Flash chat template

Delete params

Delete template

Fix: use closing tag </tool_call> for proper tool call capture

Add tool_call stop token

Simplify to single stop token like unsloth

Add num_predict -1 for unlimited output

Add num_ctx 202752 to match cloud model

Remove observation stop token to fix cutoff issue

Fix: use JSON format for tool calls (Ollama compatible)

Update params with all stop tokens

Update template to match official zai-org format

Fix stop tokens - remove unnecessary observation token

Fix GGUF file_type metadata (BF16->IQ4_XS)

Add Ollama params with stop tokens

Add Ollama template with tools and thinking support

Upload mmproj-Elbaz-GLM-4.6V-Flash-PRISM-F16.gguf with huggingface_hub

Upload Elbaz-GLM-4.6V-Flash-PRISM-IQ4_XS.gguf with huggingface_hub

Delete Elbaz-GLM-4.6V-Flash-PRISM-mmproj-bf16.gguf with huggingface_hub

Delete Elbaz-GLM-4.6V-Flash-PRISM-IQ4_XS.gguf with huggingface_hub

Upload Elbaz-GLM-4.6V-Flash-PRISM-mmproj-bf16.gguf with huggingface_hub

Upload Elbaz-GLM-4.6V-Flash-PRISM-IQ4_XS.gguf with huggingface_hub

Update README.md

Update README.md

Update README.md

Upload README.md with huggingface_hub

initial commit