baalajimaestro/ollama
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

try next runner if load fails

Browse source
This commit is contained in:
Bruce MacDonald 2023-07-04 11:55:10 -04:00
parent a86a4f8c26
commit 27ead750c3
1 changed files with 11 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

12
ollama/engine.py
View file

@ -39,12 +39,18 @@ def load(model_name, models={}):
for model_type in cls.model_types() for model_type in cls.model_types()
} }
while len(runners) > 0:
try:
best_match, _ = process.extractOne(model_path, runners.keys()) best_match, _ = process.extractOne(model_path, runners.keys())
model = runners.get(best_match, LlamaCppRunner) model = runners.get(best_match, LlamaCppRunner)
runner = model(model_path, best_match)
models.update({model_name: model(model_path, best_match)}) models.update({model_name: runner})
return models.get(model_name) return models.get(model_name)
except Exception:
# try the next runner
runners.pop(best_match)
raise Exception("failed to load model", model_path, model_name)
def unload(model_name, models={}): def unload(model_name, models={}):