docs: pipeline_tag image-text-to-text (was any-to-any)

HF's llama.cpp/Ollama manifest endpoint returned 'Repository is not GGUF
or is not compatible with llama.cpp' when pulling new BF16 and Q8_0 tags
with pipeline_tag: any-to-any. Switching to image-text-to-text (which
works on the sibling lthn/lemer repo) fixes the routing. Also adds
llama.cpp, ollama, multimodal tags for discoverability.

Co-Authored-By: Virgil <virgil@lethean.io>

README.md

@@ -2,13 +2,16 @@
 library_name: gguf
 license: apache-2.0
 license_link: https://ai.google.dev/gemma/docs/gemma_4_license
-pipeline_tag: any-to-any
+pipeline_tag: image-text-to-text
 base_model:
   - google/gemma-4-E2B-it
 tags:
   - gemma4
   - gguf
   - safetensors
+  - llama.cpp
+  - ollama
+  - multimodal
 ---
 
 <div align="center">