|
--- |
|
library_name: transformers |
|
tags: [] |
|
--- |
|
|
|
# Model Card for Model ID |
|
|
|
<!-- Provide a quick summary of what the model is/does. --> |
|
DeepAutoAI/Explore_Llama-3.1-8B-Inst is a customized variant of Llama-2.1-8B-Instruct. This customization is achieved by learning |
|
the distribution of all normalization layer weights followed by the distribution of the last transformer block, 30, and 24th FFN layers of |
|
the original Llama model. |
|
A layer-conditional diffusion based weights generation model that enables sampling for performance enhancement by leveraging |
|
the learned distributions to optimize the merging process is used to generate newly diverse weights |
|
|
|
|
|
|
|
## Model Details |
|
|
|
### Model Description |
|
|
|
<!-- Provide a longer summary of what this model is. --> |
|
|
|
We trained a diffusion model to learn the distribution of subset of llama to enable generation weights that improve the performance. |
|
We generate task specific weights on winogrande and arc_challenge then transfer the best model for leaderboard benchmarking. |
|
|
|
- **Developed by:** DeepAuto.ai |
|
- **Funded by [optional]:** DeepAuto.ai |
|
- **Shared by [optional]:** DeepAuto.ai |
|
- **Model type:** llama-3.1-8B |
|
- **Language(s) (NLP):** English |
|
- **License:** NA |
|
- **Finetuned from model [optional]:** No fine-tuning |
|
|
|
### Model Sources [optional] |
|
|
|
<!-- Provide the basic links for the model. --> |
|
|
|
- **Repository:** Under construction |
|
- **Paper [optional]:** To be announce |
|
|
|
|
|
## Uses |
|
|
|
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> |
|
|
|
|
|
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> |
|
|
|
The direct use case of our work is o improve existing model performance as well as generating task specific weights with no training. |
|
|
|
|
|
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> |
|
Performance improvement of existing large models with limited compute |
|
|
|
### Out-of-Scope Use |
|
|
|
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> |
|
|
|
No fine-tuning or architecture generalization |
|
|
|
## Bias, Risks, and Limitations |
|
|
|
<!-- This section is meant to convey both technical and sociotechnical limitations. --> |
|
|
|
Using a generative model to produce weights can potentially lead to unintended or undesirable outputs. However, the generated content |
|
will still fall within the range of what the base model is inherently capable of producing. |
|
|
|
### Recommendations |
|
|
|
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> |
|
|
|
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. |
|
|
|
## How to Get Started with the Model |
|
|
|
Use the code below to get started with the model. |
|
|
|
[More Information Needed] |
|
|
|
## Training Details |
|
We employed a latent diffusion process on pretrained model weights, unlocking the ability to generate diverse, previously unseen neural networks. |
|
Remarkably, even within the constraints of one-shot learning, our approach consistently produces a wide range of weight variations, each offering |
|
distinct performance characteristics. These generated weights not only open opportunities for weight averaging and model merging but also have the |
|
potential to significantly enhance model performance. Moreover, they enable the creation of task-specific weights, tailored to optimize performance |
|
for specialized applications |
|
|
|
### Training Data |
|
The training data used to produced the current model is the base pretrained weights |
|
|
|
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> |
|
|
|
|
|
### Training Procedure |
|
|
|
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> |
|
|
|
- We selected a set of layers and combined their pretrained weights, then trained a Variational Autoencoder (VAE) to encode these weights into the layer dimension. |
|
- We conditionally trained a diffusion model on this set of weights, allowing individual sampling of layer-specific weights. |
|
- All selected layers were encoded into a 1024-dimensional space. This model exclusively contained the sampled weights for layer normalization." |
|
|
|
|
|
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> |
|
|
|
|
|
## Evaluation |
|
|
|
<!-- This section describes the evaluation protocols and provides the results. --> |
|
|
|
### Testing Data, Factors & Metrics |
|
|
|
|
|
<!-- This should link to a Dataset Card if possible. --> |
|
|
|
We test our method on Winogrande and arc_challenge, and hellaswag |
|
|
|
#### Factors |
|
|
|
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> |
|
|
|
[More Information Needed] |
|
|
|
#### Metrics |
|
|
|
<!-- These are the evaluation metrics being used, ideally with a description of why. --> |
|
|
|
[More Information Needed] |
|
|
|
### Results |
|
|
|
[More Information Needed] |
|
|
|
#### Summary |
|
|
|
|
|
|
|
## Model Examination [optional] |
|
|
|
<!-- Relevant interpretability work for the model goes here --> |
|
|
|
[More Information Needed] |
|
|
|
## Environmental Impact |
|
|
|
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> |
|
|
|
|
|
|
|
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). |
|
|
|
- **Hardware Type:** Nvidia-A100-80Gb |
|
- **Hours used:** VAE is trained for 4 hour and diffusion process 4 hours |
|
- **Compute Region:** South Korea |
|
- **Carbon Emitted:** 0.96kg |
|
|
|
## Technical Specifications [optional] |
|
|
|
### Model Architecture and Objective |
|
|
|
We used Latent diffusion for weights generation, and llama3-1-8B as target architectures. |
|
|
|
The primary objective of this weight generation process was to demonstrate that by learning only the distribution |
|
of few layers weights9normlaization layers in this case) in an 8-billion-parameter model, it is possible to significantly enhance the |
|
model's capabilities. Notably, this is achieved using a fraction of the computational resources and without the |
|
need for fine-tuning, showcasing the efficiency and potential of this approach. |
|
|
|
### Compute Infrastructure |
|
|
|
Nvidia-A100 cluster |
|
|
|
#### Hardware |
|
|
|
A single Nvidia-A100 |
|
|
|
#### Software |
|
|
|
Model is tested using lm-harness tool version 0.4.3 |
|
|
|
## Citation [optional] |
|
|
|
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> |
|
|
|
**BibTeX:** |
|
|
|
[More Information Needed] |
|
|
|
**APA:** |
|
|
|
[More Information Needed] |
|
|
|
## Glossary [optional] |
|
|
|
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> |
|
|
|
[More Information Needed] |
|
|
|
## More Information [optional] |
|
|
|
[More Information Needed] |
|
|
|
## Model Card Authors [optional] |
|
|
|
[More Information Needed] |
|
|
|
## Model Card Contact |
|
|
|
For any question contact deepauto.ai |
