Connecting to open-source models in LangChain - Step-by-Step Execution

Concept Flow - Connecting to open-source models

Start

↓

Import LangChain

↓

Choose Open-Source Model

↓

Set Model Parameters

↓

Create Model Instance

↓

Send Input to Model

↓

Receive Output

↓

Use Output in Application

↓

End

This flow shows how to connect to an open-source model using LangChain: import, select model, set parameters, create instance, send input, get output, then use it.

Execution Sample

LangChain

from langchain.llms import HuggingFaceHub

model = HuggingFaceHub(repo_id="google/flan-t5-small")
response = model("What is AI?")
print(response)

This code connects to an open-source model on HuggingFace Hub, sends a question, and prints the answer.

Execution Table

Step	Action	Input/Parameters	Result/Output
1	Import HuggingFaceHub from langchain.llms	None	Module ready to use
2	Create model instance	repo_id="google/flan-t5-small"	Model object created
3	Send input to model	"What is AI?"	Model processes input
4	Receive output	Model generates answer	"AI is the simulation of human intelligence by machines."
5	Print output	Output from model	Printed answer on screen

💡 Execution stops after printing the model's response.

Variable Tracker

Variable	Start	After Step 2	After Step 3	After Step 4	Final
model	None	HuggingFaceHub instance	HuggingFaceHub instance	HuggingFaceHub instance	HuggingFaceHub instance
response	None	None	None	"AI is the simulation of human intelligence by machines."	"AI is the simulation of human intelligence by machines."

Key Moments - 3 Insights

Why do we need to specify repo_id when creating the model?

The repo_id tells LangChain which open-source model to load from HuggingFace Hub, as shown in step 2 of the execution_table.

What happens if the model takes time to respond?

The code waits at step 3 until the model finishes processing and returns output at step 4, so the response variable updates only after completion.

Can we reuse the model variable for multiple inputs?

Yes, the model instance stays active after step 2 and can be called multiple times with different inputs without recreating it.

Visual Quiz - 3 Questions

Look at the execution_table, what is the value of 'response' after step 3?

A None

B Processing input

C "AI is the simulation of human intelligence by machines."

D Model object created

💡 Hint

Check the variable_tracker row for 'response' after step 3.

At which step does the model instance get created?

A Step 1

B Step 2

C Step 3

D Step 4

💡 Hint

Look at the 'Action' column in execution_table for model creation.

If you want to ask another question, what should you do with the 'model' variable?

A Create a new model instance again

B Set repo_id again

C Call the model variable with the new input

D Print the model variable

💡 Hint

Refer to the key_moments about reusing the model instance.

Concept Snapshot

Connecting to open-source models with LangChain:
- Import the model class (e.g., HuggingFaceHub)
- Specify the model repo_id from HuggingFace Hub
- Create a model instance
- Call the model with input text
- Receive and use the output
- Reuse the model instance for multiple queries

Full Transcript

This lesson shows how to connect to open-source models using LangChain. First, you import the HuggingFaceHub class. Then you create a model instance by specifying the repo_id of the model you want to use. Next, you send input text to the model instance, which processes it and returns an output. Finally, you print or use the output in your application. The model instance can be reused for multiple inputs without recreating it. This step-by-step flow helps beginners see how the connection and data flow happen in code.