User Tools

Site Tools


it:ai_delphi

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
it:ai_delphi [2026/08/17 13:06] – [open source inference engines] gary1it:ai_delphi [2026/08/17 13:31] (current) – [open source inference engines] gary1
Line 32: Line 32:
     * [[https://git.rosseaux.net/BeRo1985/pasllm]]     * [[https://git.rosseaux.net/BeRo1985/pasllm]]
     * currently only uses CPU not GPU (May 2026)     * currently only uses CPU not GPU (May 2026)
-    * on my laptop with 64Gb RAM, and using a Euromoe 2.6b 1.45Gb model, and the VCL Forms version, output was 36 tokens/sec but output quality was not great with this model+    * on my laptop with 64Gb RAM,  and the VCL Forms version: 
 +      * using a Euromoe 2.6b 1.45Gb model, output was 36 tokens/sec but output quality was not great with this model 
 +      * using a Llama 3.1 8b instruct model, output was 8 tokens/sec but much better output quality
   * **Vindex Lm**   * **Vindex Lm**
     * [[https://github.com/tinyBigGAMES/VindexLLM]]     * [[https://github.com/tinyBigGAMES/VindexLLM]]
it/ai_delphi.txt · Last modified: 2026/08/17 13:31 by gary1

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki