Create README.md

Here a view embedding models that i have collected from hugging.
All file names named like or very similar to original if you like to search for.

i tested all models local with ALLM (AnythingLLM) with LM as Server
(they work, but every setting is different)

my short impression is:
- mug-b-1.6-q8_0.gguf
- mxbai-embed-large-v1.Q8_0.gguf
- nomic_f16.gguf/nomic-embed-text-v1.5.Q8_0.gguf
are okay
test other models are up to you...


short hint, set (Max Tokens)context-length of your main-model to 16k, set embedding (Max Embedding Chunk Length)1024, and (Max Context Snippets) 14
so you can receive 14 snippets a 1024token ~12000words from you document and have 2048token ~1500word left for the answer.
you can play for you need and VRAM usage



(ALLE licenses and terms of use go to original authors)

README.md

@@ -0,0 +1,13 @@
+---
+library_name: sentence-transformers
+pipeline_tag: sentence-similarity
+tags:
+  - sentence-transformers
+  - sentence-similarity
+  - feature-extraction
+  - embedder
+  - embedding
+  - moedels
+  - GGUF
+
+---