Update README.md Aquila2.

[ftgreat]

Recently we released Aquila2 models. Models are opensourced in HuggingFace, like AquilaChat2-34B

Could we update README.md with Aquila2,thanks.

[WoosukKwon]

LGTM. Thanks for the updates!

[huyz-git]

I failed to run Aquila2 with vllm:

ValueError: Model architectures ['AquilaForCausalLM'] are not supported for now. Supported architectures: ['AquilaModel', 'BaiChuanForCausalLM', 'BaichuanForCausalLM', 'BloomForCausalLM', 'FalconForCausalLM', 'GPT2LMHeadModel', 'GPTBigCodeForCausalLM', 'GPTJForCausalLM', 'GPTNeoXForCausalLM', 'InternLMForCausalLM', 'LlamaForCausalLM', 'LLaMAForCausalLM', 'MPTForCausalLM', 'OPTForCausalLM', 'QWenLMHeadModel', 'RWForCausalLM']

If I modify vllm/model_executor/model_loader.py file and map this model architechture into AquilaForCausalLM, then I got:

File "/home/xxx/vllm/vllm/model_executor/models/aquila.py", line 370, in load_weights
    assert param_slice.shape == loaded_weight.shape
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[lu-wang-dl]

Make a PR to fix the issue: #1339

[lonngxiang]

I failed to run Aquila2 with vllm:

ValueError: Model architectures ['AquilaForCausalLM'] are not supported for now. Supported architectures: ['AquilaModel', 'BaiChuanForCausalLM', 'BaichuanForCausalLM', 'BloomForCausalLM', 'FalconForCausalLM', 'GPT2LMHeadModel', 'GPTBigCodeForCausalLM', 'GPTJForCausalLM', 'GPTNeoXForCausalLM', 'InternLMForCausalLM', 'LlamaForCausalLM', 'LLaMAForCausalLM', 'MPTForCausalLM', 'OPTForCausalLM', 'QWenLMHeadModel', 'RWForCausalLM']

If I modify vllm/model_executor/model_loader.py file and map this model architechture into AquilaForCausalLM, then I got:

File "/home/xxx/vllm/vllm/model_executor/models/aquila.py", line 370, in load_weights
    assert param_slice.shape == loaded_weight.shape
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

same error;version 0.2.0

ValueError: Model architectures ['AquilaForCausalLM'] are not supported for now. Supported architectures: ['AquilaModel', 'BaiChuanForCausalLM', 'BaichuanForCausalLM', 'BloomForCausalLM', 'FalconForCausalLM', 'GPT2LMHeadModel', 'GPTBigCodeForCausalLM', 'GPTJForCausalLM', 'GPTNeoXForCausalLM', 'InternLMForCausalLM', 'LlamaForCausalLM', 'LLaMAForCausalLM', 'MPTForCausalLM', 'OPTForCausalLM', 'QWenLMHeadModel', 'RWForCausalLM']