#
Tag: tutorial
See all tags.
This tutorial is for anyone who wants to fine-tune powerful large language models such as Llama2, Mistral for their own projects.
This tutorial introduces an example of fine-tuning the open-source Llama3-8B model on the MoAI Platform.
Setting up the PyTorch execution environment on the MoAI Platform is similar to setting it up on a typical GPU server.
Once you have prepared all the training data, let's take a look at the contents of the train_llama3.py
Now, we will actually execute the fine-tuning process.
When you execute the train_llama3.py
script as in the previous section, the resulting model will be saved in the
Let's rerun the fine-tuning task with a different number of GPUs.
From this tutorial, we have seen how to fine-tune Llama3 8B on the MoAI Platform.
This tutorial introduces how to fine-tune the open-source LLama3-70B model using the MoAI Platform.
To fine-tune the Llama3 70B model, you must use multiple GPUs and implement parallelization techniques such as Tensor Parallelism, Pipeline...
For a smooth tutorial experience, the following specifications are recommended:
Now, we will actually execute the fine-tuning process.
We have now explored the process of fine-tuning the Llama3-70b model on the MoAI Platform.
This tutorial guides you on fine-tuning the open-source Mistral 7B model on the MoAI Platform.
Preparing the PyTorch script execution environment on the MoAI Platform is similar to doing so on a typical GPU server.
Once you have prepared all the training data, let's delve into the contents of the train_mistral.py
script to execute the actual fine-tuning process.
Now, we will train the model through the following process.
Running the train_mistral.py
script, as in the previous section, will save the resulting model in the
Let's rerun the fine-tuning task with a different number of GPUs.
From this tutorial, we have seen how to fine-tune the Mistral 7B model on the MoAI Platform.
This tutorial guides you on how to fine-tune GPT-based models open-sourced by Hugging Face on the MoAI Platform.
Preparing the PyTorch script execution environment on the MoAI Platform is similar to doing so on a typical GPU server.
Once you have prepared all the training data, let's take a look at the contents of the train_gpt.py
script to execute the actual fine-tuning process.
Now, we will train the model through the following process.
As in the previous chapter, when you run the train_gpt.py
script, the resulting model will be saved in the
Let's rerun the fine-tuning task with a different number of GPUs.
So far, we have looked at the process of fine-tuning the GPT-based model from HuggingFace on the MoAI Platform.
This tutorial introduces an example of fine-tuning the open-source Qwen1.5 7B model on the MoAI Platform.
Preparing the PyTorch script execution environment on the MoAI Platform is similar to doing so on a typical GPU server.
If you've prepared all the training data, let's now take a look at the train_qwen.py
script to actually run the fine-tuning process.
Now, we will train the model through the following process.
Running the train_qwen.py
script, as in the previous chapter, will save the resulting model in the qwen_code_generation
Let's rerun the fine-tuning task with a different number of GPUs.
So far, we've explored the process of fine-tuning the Qwen1.5 7B model on the MoAI Platform.
The following tutorial will take you through the steps required to fine-tune Baichuan2 13B model with an example dataset, using the MoAI Platform.
Preparing the PyTorch script execution environment on the MoAI Platform is similar to doing so on a typical GPU server.
If you've prepared all the training data, let's now take a look at the contents of train_baichuan2_13b.py
Now, we will train the model through the following process.
Similar to the previous chapter, when you execute the train_baichuan2_13b.py
script, the resulting model will be saved in the
Let's rerun the fine-tuning task with a different number of GPUs.
So far, we've explored the process of fine-tuning the Baichuan2 13B model, which is publicly available on Hugging Face, using the MoAI Platform.
This tutorial introduces an example of fine-tuning the open-source Llama2 13B model on the MoAI Platform.
Preparing the PyTorch script execution environment on the MoAI Platform is similar to doing so on a typical GPU server.
If you've got all your training data ready, let's dive into running the actual fine-tuning process using the
Now, we will train the model through the following process.
Upon running the train_llama2.py
script as described earlier, the resulting model will be saved in the
Let's rerun the fine-tuning task with a different number of GPUs.
From this tutorial, we have seen how to fine-tune Llama2 13B for text summarization on the MoAI Platform.