cleanup, up ver (+1 squashed commits)

Squashed commits:

[1ea303d] cleanup , up ver (+1 squashed commits)

Squashed commits:

[79f09b2] cleanup

colab.ipynb

@@ -67,7 +67,7 @@
         "!test -f koboldcpp_cublas.so && echo Build Skipped || make koboldcpp_cublas LLAMA_CUBLAS=1\r\n",
         "!cp koboldcpp_cublas.so koboldcpp_cublas.dat\r\n",
         "!apt install aria2 -y\r\n",
-        "!aria2c -x 10 -o model.ggml --summary-interval=3 --download-result=full --allow-overwrite=true --file-allocation=none $Model\r\n",
+        "!aria2c -x 10 -o model.ggml --summary-interval=5 --download-result=default --allow-overwrite=true --file-allocation=none $Model\r\n",
         "!python koboldcpp.py model.ggml --usecublas 0 mmq --multiuser --gpulayers $Layers --contextsize $ContextSize --hordeconfig concedo 1 1 --remotetunnel\r\n"
       ]
     }

koboldcpp.py

@@ -379,7 +379,7 @@ def bring_terminal_to_foreground():
 modelbusy = threading.Lock()
 requestsinqueue = 0
 defaultport = 5001
-KcppVersion = "1.48"
+KcppVersion = "1.48.1"
 showdebug = True
 showsamplerwarning = True
 showmaxctxwarning = True
@@ -1140,21 +1140,18 @@ def auto_gpu_heuristics():
                     MaxMemory[0] = max(int(FetchedCUdeviceMem[idx]),MaxMemory[0])
                 else:
                     MaxMemory[0] = max(int(FetchedCUdeviceMem[idx])*1024*1024,MaxMemory[0])
-                    pass
 
-        #autopick cublas if suitable
+        #autopick cublas if suitable, requires at least 3.5GB VRAM to auto pick
         global exitcounter
-        if exitcounter < 100 and MaxMemory[0]>3500000000 and CUDevicesNames[0]!="" and "Use CuBLAS" or "Use hipBLAS (ROCM)" in runopts and runopts_var.get()=="Use OpenBLAS":
+        if exitcounter < 100 and MaxMemory[0]>3500000000 and CUDevicesNames[0]!="" and ("Use CuBLAS" in runopts or "Use hipBLAS (ROCM)" in runopts) and runopts_var.get()=="Use OpenBLAS":
             if "Use CuBLAS" in runopts:
                 runopts_var.set("Use CuBLAS")
-                pass
             elif "Use hipBLAS (ROCM)" in runopts:
                 runopts_var.set("Use hipBLAS (ROCM)")
 
         changed_gpu_choice_var()
         return
 
-
     def autoset_gpu_layers(filepath): #shitty algo to determine how many layers to use
         try:
             global gui_layers_untouched