采样过程问题

[huxiaosheng123]

Describe the bug
swift sample --model /train_model/Qwen2.5-7B-Instruct/ --sampler_engine lmdeploy --num_return_sequences 5 --n_best_to_keep 3 --dataset /ms-swift/dataset/AI-ModelScope/alpaca-gpt4-data-zh#5 --prm_model /train_model/GRM-Llama3.2-3B-rewardmodel-ft/

swift sample --model /train_model/Qwen2.5-7B-Instruct/ --sampler_engine lmdeploy --num_return_sequences 5 --n_best_to_keep 3 --dataset /ms-swift/dataset/AI-ModelScope/alpaca-gpt4-data-zh#5 --orm_model math

都报错 /ms-swift/swift/llm/sampling/vanilla_sampler.py", line 136, in generate
if resp_list[j].choices[0].message.content:
AttributeError: 'AttributeError' object has no attribute 'choices'

--sampler_engine lmdeploy 换成pt，报错：'PtEngine' object is not callable

[tastelikefeet]

拉下最新的main代码，已经修复了 #

[huxiaosheng123]

拉下最新的main代码，已经修复了 #

Traceback (most recent call last):
File "/ms-swift-main/swift/cli/sample.py", line 5, in
sampling_main()
File "/ms-swift-main/swift/llm/sampling/sampling.py", line 73, in sampling_main
return SwiftSampling(args).main()
File "/ms-swift-main/swift/llm/base.py", line 46, in main
result = self.run()
File "/ms-swift-main/swift/llm/sampling/sampling.py", line 64, in run
generated = self.sampler.do_sample(slices)
File "/ms-swift-main/swift/llm/sampling/vanilla_sampler.py", line 169, in do_sample
prm_score, _prm_mask = get_reward(
File "/ms-swift-main/swift/llm/sampling/utils.py", line 45, in get_reward
rewards = model(infer_requests, request_config=request_config, **gt_param)
TypeError: 'PtEngine' object is not callable

刚拉的main代码，还是会报错

--sampler_engine lmdeploy 方式：
AttributeError("'AsyncEngine' object has no attribute 'get_generator'")

[tastelikefeet]

抱歉最近我们在忙于GRPO的事情，已修复 #3184