AirLLM 2.8T LLM inference with single 4GB GPU
Python chinese-llmchinese-nlpfinetunegenerative-aiinstruct-gptinstruction-setllamallmloraopen-modelsopen-sourceopen-source-modelsqlora
fix test
Yu Lib5cd55ff4463
diff --git a/air_llm/airllm/auto_model.py b/air_llm/airllm/auto_model.pyindex e91744c..c6c4bc7 100644--- a/air_llm/airllm/auto_model.py+++ b/air_llm/airllm/auto_model.py@@ -17,18 +17,18 @@ class AutoModel:@classmethoddef from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs):config = AutoConfig.from_pretrained(pretrained_model_name_or_path, trust_remote_code=True)- if "QWen" in config['architectures']:+ if "QWen" in config.architectures[0]:return AirLLMQWen(*inputs, **kwargs)- elif "Baichuan" in config['architectures']:+ elif "Baichuan" in config.architectures[0]:return AirLLMBaichuan(*inputs, **kwargs)- elif "ChatGLM" in config['architectures']:+ elif "ChatGLM" in config.architectures[0]:return AirLLMChatGLM(*inputs, **kwargs)- elif "InternLM" in config['architectures']:+ elif "InternLM" in config.architectures[0]:return AirLLMInternLM(*inputs, **kwargs)- elif "Mistral" in config['architectures']:+ elif "Mistral" in config.architectures[0]:return AirLLMMistral(*inputs, **kwargs)- elif "Llama" in config['architectures']:+ elif "Llama" in config.architectures[0]:return AirLLMLlama2(*inputs, **kwargs)else:- print(f"unknown artichitecture: {config['architectures']}, try to use Llama2...")+ print(f"unknown artichitecture: {config.architectures[0]}, try to use Llama2...")return AirLLMLlama2(*inputs, **kwargs)\ No newline at end of file