deepseek-ai

/

DeepSeek-Coder-V2-Lite-Instruct

like 393

Follow

DeepSeek 38.8k

File "/root/.cache/huggingface/modules/transformers_modules/DeepSeek-Coder-V2-Lite-Instruct/modeling_deepseek.py", line 585, in forward
y = self.moe_infer(hidden_states, topk_idx, topk_weight).view(*orig_shape)
RuntimeError: shape '[1, 1, 2048]' is invalid for input of size 0