How to use zai-org/chatglm2-6b with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("zai-org/chatglm2-6b", trust_remote_code=True, dtype="auto")
Hi~ Can anyone know the multi_query_group_num? Please help to understand the meaning of it."multi_query_attention": true, "multi_query_group_num": 2, "num_attention_heads": 32, "num_layers": 28,
· Sign up or log in to comment