Some Weights in GemmaForSequenceClassification Are Not Initialized from Checkpoint
#6
by
KevinWangHP
- opened
Some weights of GemmaForSequenceClassification were not initialized from the model checkpoint at ../pretrained_model/BAAI/bge-reranker-v2-gemma and are newly initialized: ['score.weight'] You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. Why is score.weight not initialized from the checkpoint?
self.model = AutoModelForSequenceClassification.from_pretrained(model_path).half().to(device).eval()
KevinWangHP
changed discussion status to
closed