Update README.md

README.md

@@ -9,13 +9,18 @@ This is an experimental coding-focused merge of the latest of two of my favorite
 Replete-AI's Replete LLM Qwen2-7B (https://huggingface.co/Replete-AI/Replete-LLM-Qwen2-7b)
 Arcee-AI's Arcee Spark (https://huggingface.co/arcee-ai/Arcee-Spark)
 
-If there is interest, I will publish the huggingface repo; otherwise, the dare_ties mergekit YAML that produced it is already included.
+If there is interest, I will publish the huggingface repo; otherwise, the dare_ties mergekit YAML that produced it is already included.  The GGUF is quantized to q8_0 for output and embedding tensors, q5_k_m for others.
 
 I asked it to describe itself, given its own mergekit YAML and points from the web pages for its two ancestor models:
 
-```This is the mergekit YAML that made you from an experimental DARE TIES merge between Arcee Spark and Replete LLM Qwen2 7b. This is an effort to combine the strengths of two Qwen2-based models which train and fine-tune the model on high-quality open source data.
-Describe what the mergekit YAML implies, then write a paragraph about what the features of this model might be considering that the density favors the coding-oriented Replete LLM Qwen2, and supply a bullet list of the ten top likely features of such a merge. Then, write a sentence about the impact of a GGUF quantization which uses q8_0 for output and embedding tensors, and q5_k_m for the rest.```
+```
+This is the mergekit YAML that made you from an experimental DARE TIES merge between Arcee Spark and Replete LLM Qwen2 7b. This is an effort to combine the strengths of two Qwen2-based models which train and fine-tune the model on high-quality open source data.
+Describe what the mergekit YAML implies, then write a paragraph about what the features of this model might be considering that the density favors the coding-oriented Replete LLM Qwen2, and supply a bullet list of the ten top likely features of such a merge. Then, write a sentence about the impact of a GGUF quantization which uses q8_0 for output and embedding tensors, and q5_k_m for the rest.
+```
 
+Here's the response:
+
+```
 The mergekit YAML described here suggests an intricate process that amalgamates two models - Arcee Spark and Replete AI's LLM Qwen2-7b, both initialized from the base model Qwen/Qwen2-7B. The merging strategy is based on the DARE TIES method, which typically aims to combine models in a way that leverages their individual strengths while minimizing any potential weaknesses.
 
 Considering the density and weight parameters favor the Replete LLM Qwen2, it implies that this model might have been fine-tuned or optimized for tasks related to coding. The Replete AI's model is known for its versatility and performance on various types of tasks, making it a valuable addition when merged with Arcee Spark.
@@ -34,4 +39,5 @@ Ten likely features of such a merge could include:
   * Improved Function Calling Capabilities
 
 The impact of using GGUF quantization with q8_0 for output and embedding tensors, and q5_k_m for the rest would be significant. This form of quantization reduces the precision of numerical values in the model, which can lead to a substantial decrease in memory usage and computational requirements without significantly compromising performance. The trade-off might involve minor loss in accuracy or quality due to reduced precision but could make the model more deployable on resource-constrained devices or platforms.
-In summary, this mergekit YAML implies an advanced AI system that leverages the strengths of both Arcee Spark and Replete LLM Qwen2-7b for versatile applications while optimizing performance and efficiency.
+In summary, this mergekit YAML implies an advanced AI system that leverages the strengths of both Arcee Spark and Replete LLM Qwen2-7b for versatile applications while optimizing performance and efficiency.
+```