chat_engine = CondenseQuestionChatEngine.from_defaults(
    query_engine=index.as_query_engine(),
    condense_question_prompt=custom_prompt,
    chat_history=custom_chat_history,
    verbose=True,
    llm=llm,
    service_context=service_context,
)

KeyError: 'custom_chat_history'

import chromadb
from llama_index import VectorStoreIndex
from llama_index.vector_stores import ChromaVectorStore
from llama_index.storage.storage_context import StorageContext
from llama_index.llms import LlamaCPP
from llama_index import VectorStoreIndex, SimpleDirectoryReader, ServiceContext

# initialize client
db = chromadb.PersistentClient(path="./chroma_db")
llm = LlamaCPP(
    # You can pass in the URL to a GGML model to download it automatically
    # optionally, you can set the path to a pre-downloaded model instead of model_url
    model_path="./models/em_german_13b_v01.Q8_0.gguf",
    temperature=0.1,
    max_new_tokens=4048,
    # llama2 has a context window of 4096 tokens, but we set it lower to allow for some wiggle room
    context_window=8128,
    # kwargs to pass to __call__()
    generate_kwargs={},
    # kwargs to pass to __init__()
    # set to at least 1 to use GPU
    # model_kwargs={"n_gpu_layers": 1},
    # transform inputs into Llama2 format
    # messages_to_prompt=messages_to_prompt,
    # completion_to_prompt=completion_to_prompt,
    verbose=True,
    
)
# get collection
chroma_collection = db.get_or_create_collection("quickstart")

# assign chroma as the vector_store to the context
vector_store = ChromaVectorStore(chroma_collection=chroma_collection)
storage_context = StorageContext.from_defaults(vector_store=vector_store)
service_context = ServiceContext.from_defaults(llm=llm)

# load your index from stored vectors
index = VectorStoreIndex.from_vector_store(
    vector_store, storage_context=storage_context, service_context=service_context
)

# create a query engine
query_engine = index.as_query_engine()
response = query_engine.query("Hallo, wie geht es dir?")
print(response)