How to use Qwen/Qwen3-Embedding-8B with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("Qwen/Qwen3-Embedding-8B") sentences = [ "The weather is lovely today.", "It's so sunny outside!", "He drove to the stadium." ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3]
How to use Qwen/Qwen3-Embedding-8B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="Qwen/Qwen3-Embedding-8B")
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Qwen/Qwen3-Embedding-8B", dtype="auto")
embedding模型的embed_tokens layer不能直接用作lm_heads，输出是乱码。如果用qwen3-8b的lm_heads，也是乱码。是不是必须得重新训练一个lm_heads?如果要训练的话，请问官方大佬有什么建议呢，是freeze 主干只训练lm_heads，还是full params一起训练呢？推荐用多少数据？
· Sign up or log in to comment
