dragneel2074 commited on
Commit
d7ecf2b
1 Parent(s): fe312ba

using tinyllama

Browse files
Files changed (2) hide show
  1. app.py +1 -1
  2. start.sh +2 -2
app.py CHANGED
@@ -14,7 +14,7 @@ logging.basicConfig(level=logging.INFO)
14
  logger = logging.getLogger(__name__)
15
 
16
  app = FastAPI()
17
- MODEL_NAME = 'gemma2:2b'
18
 
19
  @lru_cache()
20
  def get_llm():
 
14
  logger = logging.getLogger(__name__)
15
 
16
  app = FastAPI()
17
+ MODEL_NAME = 'tinyllama'
18
 
19
  @lru_cache()
20
  def get_llm():
start.sh CHANGED
@@ -9,8 +9,8 @@ export CUDA_VISIBLE_DEVICES=0 # Use the first GPU if available
9
  ollama serve & # Use GPU 0 if available
10
 
11
  # Pull the model if not already present
12
- if ! ollama list | grep -q "gemma2:2b"; then
13
- ollama pull gemma2:2b
14
  fi
15
 
16
  # Wait for Ollama to start up (use a more robust check)
 
9
  ollama serve & # Use GPU 0 if available
10
 
11
  # Pull the model if not already present
12
+ if ! ollama list | grep -q "tinyllama"; then
13
+ ollama pull tinyllama
14
  fi
15
 
16
  # Wait for Ollama to start up (use a more robust check)