From 7691a755d6b763783bd751c49f3dd6c4d53dca99 Mon Sep 17 00:00:00 2001
From: Toni Bennasar <toni.bennasar@gmail.com>
Date: Tue, 15 Apr 2025 08:56:50 +0200
Subject: [PATCH] Fixed LLama.Web errors starting session

InferenceOptions.SamplingPipeline now is not required. Is not sent from client
Added options to debug SignalR calls at appsettings.Development.json
Changed default web model to llava-v1.6-mistral-7b.Q3_K_XS.gguf, previous one (llama-2-7b-chat.Q4_0.gguf) did no exist
---
 LLama.Web/Common/InferenceOptions.cs   | 2 +-
 LLama.Web/appsettings.Development.json | 4 +++-
 LLama.Web/appsettings.json             | 6 +++---
 3 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/LLama.Web/Common/InferenceOptions.cs b/LLama.Web/Common/InferenceOptions.cs
index e5735be63..c49d3aa31 100644
--- a/LLama.Web/Common/InferenceOptions.cs
+++ b/LLama.Web/Common/InferenceOptions.cs
@@ -20,6 +20,6 @@ public class InferenceOptions
         public IReadOnlyList<string> AntiPrompts { get; set; } = Array.Empty<string>();
 
         /// <inheritdoc />
-        public required ISamplingPipeline SamplingPipeline { get; set; }
+        public ISamplingPipeline SamplingPipeline { get; set; } = new DefaultSamplingPipeline();
     }
 }
diff --git a/LLama.Web/appsettings.Development.json b/LLama.Web/appsettings.Development.json
index 770d3e931..ac6d49da3 100644
--- a/LLama.Web/appsettings.Development.json
+++ b/LLama.Web/appsettings.Development.json
@@ -3,7 +3,9 @@
   "Logging": {
     "LogLevel": {
       "Default": "Information",
-      "Microsoft.AspNetCore": "Warning"
+      "Microsoft.AspNetCore": "Warning",
+      "Microsoft.AspNetCore.SignalR": "Debug",
+      "Microsoft.AspNetCore.Http.Connections": "Debug"
     }
   }
 }
diff --git a/LLama.Web/appsettings.json b/LLama.Web/appsettings.json
index a7a627d9b..caa27cc64 100644
--- a/LLama.Web/appsettings.json
+++ b/LLama.Web/appsettings.json
@@ -10,13 +10,13 @@
     "ModelLoadType": 0,
     "Models": [
       {
-        "Name": "Example LLama2-7b-Chat",
+        "Name": "Example LLava-v1.6-mistral",
         "MaxInstances": 20,
-        "ModelPath": "..\\LLama.Unittest\\Models\\llama-2-7b-chat.Q4_0.gguf",
+        "ModelPath": "..\\LLama.Unittest\\Models\\llava-v1.6-mistral-7b.Q3_K_XS.gguf",
         "ContextSize": 2048,
         "BatchSize": 2048,
         "Threads": 4,
-        "GpuLayerCount": 6,
+        "GpuLayerCount": 32,
         "UseMemorymap": true,
         "UseMemoryLock": false,
         "MainGpu": 0,