Lab for Ilmu dasar LLM dengan Langchain

In [1]:

import json

import json

In [2]:

from langchain_ollama.llms import OllamaLLM
ollamaModel = "llama3.2:latest"
model = OllamaLLM(model=ollamaModel)

from langchain_ollama.llms import OllamaLLM ollamaModel = "llama3.2:latest" model = OllamaLLM(model=ollamaModel)

In [3]:

return_value = model.invoke("Assalamualaikum")
print(return_value)

return_value = model.invoke("Assalamualaikum") print(return_value)

Wa alaykums salam. How can I assist you today?

Konversasi menggunakan chat model¶

In [4]:

from langchain_core.messages import HumanMessage, AIMessage
from langchain_ollama.chat_models import ChatOllama

chat_model = ChatOllama(model=ollamaModel)

from langchain_core.messages import HumanMessage, AIMessage from langchain_ollama.chat_models import ChatOllama chat_model = ChatOllama(model=ollamaModel)

In [5]:

human_message = HumanMessage("Nama saya farras")

return_value_chage_model = chat_model.invoke([human_message])
print(type(return_value_chage_model))
print(json.dumps(return_value_chage_model.model_dump(), indent=4))

human_message = HumanMessage("Nama saya farras") return_value_chage_model = chat_model.invoke([human_message]) print(type(return_value_chage_model)) print(json.dumps(return_value_chage_model.model_dump(), indent=4))

<class 'langchain_core.messages.ai.AIMessage'>
{
    "content": "Selamat pagi/siangat, Farras! Bagaimana kabar Anda hari ini?",
    "additional_kwargs": {},
    "response_metadata": {
        "model": "llama3.2:latest",
        "created_at": "2025-09-06T09:45:19.8095877Z",
        "done": true,
        "done_reason": "stop",
        "total_duration": 456539500,
        "load_duration": 153339100,
        "prompt_eval_count": 30,
        "prompt_eval_duration": 16168200,
        "eval_count": 21,
        "eval_duration": 284049200,
        "model_name": "llama3.2:latest"
    },
    "type": "ai",
    "name": null,
    "id": "run--af8e3ec2-f47c-444d-91a3-aede787d86c8-0",
    "example": false,
    "tool_calls": [],
    "invalid_tool_calls": [],
    "usage_metadata": {
        "input_tokens": 30,
        "output_tokens": 21,
        "total_tokens": 51
    }
}

Build simple class to using chat models¶

In [6]:

from ChatModelsSimplifier import MyLLMChatModel

my_simple_model = MyLLMChatModel("llama3.2:latest")

from ChatModelsSimplifier import MyLLMChatModel my_simple_model = MyLLMChatModel("llama3.2:latest")

In [7]:

my_simple_model.chat_models("Halo nama saya muhammad farras ma'ruf")

my_simple_model.chat_models("Halo nama saya muhammad farras ma'ruf")

Senang bertemu dengan Muhammad Farras Ma'Ruf XXX

In [8]:

my_simple_model.chat_models("Saya punya tiga anak, 2 laki-laku, dan satu perempuan")

my_simple_model.chat_models("Saya punya tiga anak, 2 laki-laku, dan satu perempuan")

Jika Anda memiliki 3 anak, maka keluarga Anda pasti sangat bahagia dengan anggota baru yang sedang berusia XXX

In [9]:

my_simple_model.chat_models("Siapa nama saya ?")

my_simple_model.chat_models("Siapa nama saya ?")

Anda bernama Muhammad Farras Ma'Ruf XXX

In [10]:

print(my_simple_model.list_of_chats)

print(my_simple_model.list_of_chats)

[SystemMessage(content='You are assitant named NoSys, you always answert the question with sufix three XXX.', additional_kwargs={}, response_metadata={}), HumanMessage(content="Halo nama saya muhammad farras ma'ruf", additional_kwargs={}, response_metadata={}), AIMessage(content="Senang bertemu dengan Muhammad Farras Ma'Ruf XXX", additional_kwargs={}, response_metadata={'model': 'llama3.2:latest', 'created_at': '2025-09-06T09:45:20.6688002Z', 'done': True, 'done_reason': 'stop', 'total_duration': 414237400, 'load_duration': 188611600, 'prompt_eval_count': 58, 'prompt_eval_duration': 19896400, 'eval_count': 14, 'eval_duration': 204529200, 'model_name': 'llama3.2:latest'}, id='run--d5df3106-3a88-4f63-b3a9-ebda48431225-0', usage_metadata={'input_tokens': 58, 'output_tokens': 14, 'total_tokens': 72}), HumanMessage(content='Saya punya tiga anak, 2 laki-laku, dan satu perempuan', additional_kwargs={}, response_metadata={}), AIMessage(content='Jika Anda memiliki 3 anak, maka keluarga Anda pasti sangat bahagia dengan anggota baru yang sedang berusia XXX', additional_kwargs={}, response_metadata={'model': 'llama3.2:latest', 'created_at': '2025-09-06T09:45:21.2900196Z', 'done': True, 'done_reason': 'stop', 'total_duration': 610637200, 'load_duration': 156224700, 'prompt_eval_count': 100, 'prompt_eval_duration': 8874500, 'eval_count': 31, 'eval_duration': 443099000, 'model_name': 'llama3.2:latest'}, id='run--2a5e83fb-815d-4f5c-9204-dcf27af7a8a7-0', usage_metadata={'input_tokens': 100, 'output_tokens': 31, 'total_tokens': 131}), HumanMessage(content='Siapa nama saya ?', additional_kwargs={}, response_metadata={}), AIMessage(content="Anda bernama Muhammad Farras Ma'Ruf XXX", additional_kwargs={}, response_metadata={'model': 'llama3.2:latest', 'created_at': '2025-09-06T09:45:21.6142554Z', 'done': True, 'done_reason': 'stop', 'total_duration': 317733900, 'load_duration': 141230500, 'prompt_eval_count': 145, 'prompt_eval_duration': 9308300, 'eval_count': 12, 'eval_duration': 165639200, 'model_name': 'llama3.2:latest'}, id='run--e97b0621-bbd0-41f9-8b64-7c74e1a0d37f-0', usage_metadata={'input_tokens': 145, 'output_tokens': 12, 'total_tokens': 157})]

Membuat MML yang dapat digunakan kembalim¶

In [ ]: