在Google Colab部署personal-assistant项目遇多类报错求助

一、retrievalQA.py模型加载失败（AssertionError） #

错误信息 #

Loading the Manticore-13B.ggmlv2.q5_1.bin model...
llama.cpp: loading model from models/manticore-13b/Manticore-13B.ggmlv2.q5_1.bin
error loading model: unknown (magic, version) combination: 4f44213c, 50595443; is this really a GGML file?
llama_init_from_file: failed to load model
Traceback (most recent call last):
File "/content/personal-assistant/retrievalQA.py", line 61, in
main()
File "/content/personal-assistant/retrievalQA.py", line 40, in main
llm = load_local_model(model_path, provider='llamacpp')
File "/content/personal-assistant/retrievalQA.py", line 28, in load_local_model
llm = LlamaLLM(model_path, n_gpu_layers=n_gpu_layers,
File "/content/personal-assistant/pa/llm/llamacpp.py", line 20, in init
self.model = Llama(model_path=model_path,
File "/usr/local/lib/python3.10/dist-packages/llama_cpp/llama.py", line 162, in init
assert self.ctx is not None
AssertionError

解决步骤 #

• 核心原因是llama-cpp-python新版本已切换到GGUF格式，不再支持旧GGMLv2/v3格式

1. 下载Manticore-13B模型的GGUF格式版本
2. 升级llama-cpp-python到最新稳定版：

   pip install --upgrade llama-cpp-python
   

3. 修改代码中模型路径为GGUF文件的路径，确保加载逻辑匹配新格式

二、旧版llama-cpp-python的参数错误（TypeError） #

错误信息 #

TypeError: Llama.init() got an unexpected keyword argument 'n_gpu_layers'

解决步骤 #

• n_gpu_layers是llama-cpp-python较新版本新增参数，旧版本（如0.1.25）不支持

1. 优先选择升级llama-cpp-python到支持该参数的版本（参考上面的升级命令），保留GPU加速能力
2. 若必须用旧版本，删除代码中n_gpu_layers参数（但Colab CPU运行13B模型会极慢）

三、inject.py的Embedding编码错误（IndexError） #

错误信息 #

load INSTRUCTOR_Transformer
max_seq_length 512
Traceback (most recent call last):
File "/content/personal-assistant/inject.py", line 66, in
main()
File "/content/personal-assistant/inject.py", line 59, in main
db = Chroma.from_documents(texts, instructor_embeddings,
File "/usr/local/lib/python3.10/dist-packages/langchain/vectorstores/chroma.py", line 435, in from_documents
return cls.from_texts(
File "/usr/local/lib/python3.10/dist-packages/langchain/vectorstores/chroma.py", line 403, in from_texts
chroma_collection.add_texts(texts=texts, metadatas=metadatas, ids=ids)
File "/usr/local/lib/python3.10/dist-packages/langchain/vectorstores/chroma.py", line 148, in add_texts
embeddings = self._embedding_function.embed_documents(list(texts))
File "/usr/local/lib/python3.10/dist-packages/langchain/embeddings/huggingface.py", line 158, in embed_documents
embeddings = self.client.encode(instruction_pairs, **self.encode_kwargs)
File "/usr/local/lib/python3.10/dist-packages/InstructorEmbedding/instructor.py", line 524, in encode
if isinstance(sentences[0],list):
IndexError: list index out of range

解决步骤 #

• 核心原因是传入的texts列表为空，无有效文档可处理

1. 检查inject.py中数据源的路径、读取逻辑，确保能加载到有效文本内容
2. 在生成texts后添加空值判断，提前终止流程避免报错：

   if not texts:
       raise ValueError("No documents found to process. Check your data source.")
   

3. 确保Langchain与InstructorEmbedding版本兼容，建议安装指定版本：

   pip install langchain==0.0.300 InstructorEmbedding==1.0.1
   

内容的提问来源于stack exchange，提问作者eclipt