添加--quantization_bit 8 全量微调后 加载的模型32K 模型和checkpoint不对应,导致推理失败 #559
Closed
15024287710Jackson
started this conversation in
Bad Case
Replies: 2 comments
-
执行的指令: |
Beta Was this translation helpful? Give feedback.
0 replies
-
#253 这个错误我们暂时没有办法复现,你可以去这个讨论区问问社群的人 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
(ChatGLM3EnvTest) [ma-user finetune_chatmodel_demo]$python inference.py --pt-checkpoint "/home/ma-user/work/output2023120702_finetune_ds_1000data/govern_1000_data-20231207-180940-1e-4/checkpoint-200" --model /home/ma-user/work/chatglm3-6b-32k
[2023-12-08 15:00:13,259] [INFO] [real_accelerator.py:158:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Loading checkpoint shards: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:06<00:00, 1.11it/s]
Some weights of ChatGLMForConditionalGeneration were not initialized from the model checkpoint at /home/ma-user/work/chatglm3-6b-32k and are newly initialized: ['transformer.prefix_encoder.embedding.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
Traceback (most recent call last):
File "inference.py", line 29, in
model.transformer.prefix_encoder.load_state_dict(new_prefix_state_dict)
File "/home/ma-user/anaconda3/envs/ChatGLM3EnvTest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1605, in load_state_dict
self.class.name, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for PrefixEncoder:
Missing key(s) in state_dict: "embedding.weight".
(ChatGLM3EnvTest) [ma-user finetune_chatmodel_demo]$
Beta Was this translation helpful? Give feedback.
All reactions