CPU Running Comamnd
====================

```shell
/home/palash/local-ai/llama.cpp/build/bin/llama-server \
  --host 0.0.0.0 \
  --port 8090 \
  -m /home/palash/local-ai-files/my-models/gemma4-e4b-qat/gemma-4-E4B-it-qat-UD-Q4_K_XL.gguf \
  --mmproj /home/palash/local-ai-files/my-models/gemma4-e4b-qat/mmproj-BF16.gguf \
  --jinja \
  --ctx-size 262144 \
  -ctk q8_0 \
  --temp 0.2 \
  --top-p 0.8 \
  --top-k 40 \
  --min-p 0.05 \
  --repeat-penalty 1.1 \
  -t 8 \
  -tb 8 \
  -ub 1024 \
  --device none \
  --timeout 3600
  ```
  -------------------------------------------------------------------------------------------------------
  
  GPU Running Command
  ====================
 
 ```shell
  /home/palash/local-ai/llama.cpp/build/bin/llama-server \
  --host 0.0.0.0 \
  --port 8090 \
  -m /home/palash/local-ai-files/my-models/gemma4-e4b-qat/gemma-4-E4B-it-qat-UD-Q4_K_XL.gguf \
  --jinja \
  -ngl 99 \
  -nkvo \
  --ctx-size 32768 \
  -ctk q8_0 \
  --temp 1.0 \
  --top-p 0.95 \
  --top-k 64 \
  --min-p 0.05 \
  -t 8 \
  -tb 8 \
  -ub 512 \
  --timeout 3600
  
 ```
 
 ```shell
 /home/palash/local-ai/llama.cpp/build/bin/llama-server \
  --host 0.0.0.0 \
  --port 8090 \
  -m /home/palash/local-ai-files/my-models/gemma4-e4b-qat/gemma-4-E4B-it-qat-UD-Q4_K_XL.gguf \
  --jinja \
  -ngl 99 \
  -fa on\
  --ctx-size 131072 \
  -ctk q8_0 \
  -ctv q8_0 \
  --temp 1.0 \
  --top-p 0.95 \
  --top-k 64 \
  --min-p 0.05 \
  -t 8 \
  -tb 8 \
  -ub 512 \
  --timeout 3600
  -np 2
  ```
  
  
  the image generation queue processes images back to back, which is causing GPU overheating