|
--- |
|
license: mit |
|
base_model: microsoft/Phi-3-mini-128k-instruct |
|
library_name: adapters |
|
datasets: |
|
- awels/druidai_admin_dataset |
|
language: |
|
- en |
|
widget: |
|
- text: Who are you, Merlin ? |
|
tags: |
|
- awels |
|
- druidai |
|
--- |
|
|
|
# Merlin Model Card |
|
|
|
## Model Details |
|
**Model Name:** Merlin |
|
|
|
**Model Type:** Transformer-based leveraging Microsoft Phi 3b 128k tokens |
|
|
|
**Publisher:** Awels Engineering |
|
|
|
**License:** MIT |
|
|
|
**Model Description:** |
|
Merlin is a sophisticated model designed to help as an AI agent focusing on the Druid AI Conversational platform. It leverages advanced machine learning techniques to provide efficient and accurate solutions. It has been trained on the full docments corpus of Druid 7.14. |
|
|
|
## Dataset |
|
**Dataset Name:** [awels/druidai_admin_dataset](https://huggingface.co/datasets/awels/druidai_admin_dataset) |
|
|
|
**Dataset Source:** Hugging Face Datasets |
|
|
|
**Dataset License:** MIT |
|
|
|
**Dataset Description:** |
|
The dataset used to train Merlin consists of all the public documents available on the Druid AI Conversational Platform. This dataset is curated to ensure a comprehensive representation of typical administrative and development scenarios encountered in Druid AI Platform. |
|
|
|
## Training Details |
|
|
|
**Training Data:** |
|
The training data includes 33,000 Questions and Answers generated by the [Bonito LLM](https://github.com/BatsResearch/bonito). The dataset is split into 3 sets of data (training, test and validation) to ensure robust model performance. |
|
|
|
**Training Procedure:** |
|
Thready was trained using supervised learning with cross-entropy loss and the Adam optimizer. The training involved 1 epoch, a batch size of 4, a learning rate of 5.0e-06, and a cosine learning rate scheduler with gradient checkpointing for memory efficiency. |
|
|
|
**Hardware:** |
|
The model was trained on a single NVIDIA RTX 4090 graphic card. |
|
|
|
**Framework:** |
|
The training was conducted using PyTorch. |
|
|
|
## Evaluation |
|
|
|
**Evaluation Metrics:** |
|
Thready was evaluated on the training dataset: |
|
|
|
> epoch = 1.0 |
|
total_flos = 33926962GF |
|
train_loss = 2.8776 |
|
train_runtime = 0:19:34.86 |
|
train_samples_per_second = 21.546 |
|
train_steps_per_second = 5.387 |
|
|
|
**Performance:** |
|
The model achieved the following results on the evaluation dataset: |
|
|
|
> epoch = 1.0 |
|
eval_loss = 2.3814 |
|
eval_runtime = 0:01:04.90 |
|
eval_samples = 5298 |
|
eval_samples_per_second = 98.718 |
|
eval_steps_per_second = 24.683 |
|
|
|
|
|
## Intended Use |
|
|
|
**Primary Use Case:** |
|
Merlin is intended to be used locally in an agent swarm to colleborate together to solve Druid AI Conversational platform related problems. |
|
|
|
**Limitations:** |
|
While Merlin is highly effective, it may have limitations due to the model size. An 8b model based on Llama 3 is used internally at Awels Engineering. |
