Made with Supademo
Create step-by-step product demos and tutorials for your company for free. No credit card required.
Efficient OpenHathi: NyunAI
Shwu
Updated: Mar 10, 2024
Description
Click through a step-by-step, interactive demo walkthrough of Nyunai, powered by Supademo.
Steps
1
Once the infrastructure is connected select the 'Models & Datasets' tab on the platform.
2
Now, to initiate the import, please find and click on the 'Import' button.
3
Tell us a model name in the displayed input field. We'll name this model as 'OpenHaathi' and move to the next input field.
4
We'll use Open Haathi as an example model from hugging face. Move to the model website and copy the model id for the same. You can also use your own models
5
Then we fill in the pretrained model input field with the copied model id.
In case of a generally available model from huggingface, this should be the id of that model.
Next, provide the framework details in the framework input field.
6
For language models, we can directly utilise the hugging face framework.
7
Move your focus to the 'Import Model' button and give it a click.
8
Let's now import a dataset. Click on 'Datasets'
9
To continue the import process, click on the 'Import' button.
10
Kindly enter your datasets name in the text box appearing next.
11
Please choose your dataset type in the subsequent field.
12
Now, select the 'LLM - Multi Columns' option.
13
Afterward, find and select the option to 'Import Dataset.'
14
Let's move back to 'Models' by clicking on it.
15
Select 'Optimise Model' to start the optimization of your model.
16
Confirm the Optimization technique by clicking on the displayed option.
17
Now, continue by clicking 'Continue'.
18
Select your model from the drop-down option in the next step.
19
Next, select a model from the list of imported models, 'OpenHathi' in our case.
20
Please select your preferred dataset from the next field.
21
next selecting the 'samvaad' dataset that we imported
22
Go ahead by choosing a processing method in the next field.
23
Next, select the 'LLM Engine TensorRT' option that shows up on the screen.
24
Now pick a task by filling out your choice in the next input box.
25
Move ahead by selecting 'LLM' from the given options.
26
Now, enter the required parameters in the text area that appears.
27
Click on 'Kompress Model' to proceed further with your model.
28
You can now view the progress by clicking on 'In-Progress 1'.
29
To see the logs of the process, click on 'View Details'.