|
|
--- |
|
|
license: mit |
|
|
language: |
|
|
- en |
|
|
base_model: |
|
|
- gpt-4o-2024-08-06-codette |
|
|
- Raiff1982/coder |
|
|
- Raiff1982/Codette |
|
|
library_name: adapter-transformers |
|
|
datasets: |
|
|
- Raiff1982/coredata |
|
|
- Raiff1982/pineco |
|
|
metrics: |
|
|
- code_eval |
|
|
- bleurt |
|
|
- bleu |
|
|
- accuracy |
|
|
- bertscore |
|
|
- brier_score |
|
|
tags: |
|
|
- code |
|
|
- chemistry |
|
|
- legal |
|
|
- climate |
|
|
pipeline_tag: question-answering |
|
|
new_version: Raiff1982/deepercodette |
|
|
--- |
|
|
|
|
|
# Model Card for Model ID |
|
|
|
|
|
<!-- Provide a quick summary of what the model is/does. --> |
|
|
|
|
|
This model card aims to be a base template for new models. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
### Model Description |
|
|
|
|
|
<!-- Provide a longer summary of what this model is. --> |
|
|
|
|
|
This model is designed for question-answering tasks and has been fine-tuned from several base models to enhance its performance and usability. It leverages datasets from various sources to improve its accuracy and robustness. |
|
|
|
|
|
- **Developed by:** [Jonathan Harrison](https://www.office.com/search?q=Jonathan+Harrison&EntityRepresentationId=cbf3097b-72bf-4444-952d-1e473728191f) |
|
|
- **Funded by [optional]:** [More Information Needed] |
|
|
- **Shared by [optional]:** [More Information Needed] |
|
|
- **Model type:** Question-Answering |
|
|
- **Language(s) (NLP):** English |
|
|
- **License:** MIT |
|
|
- **Finetuned from model [optional]:** deepseek-ai/DeepSeek-V3 |
|
|
|
|
|
### Model Sources |
|
|
|
|
|
<!-- Provide the basic links for the model. --> |
|
|
|
|
|
- **Repository:** The model's code and configuration files can be found in the readme |
|
|
- **Paper [optional]:** [More Information Needed] |
|
|
- **Demo [optional]:** |
|
|
|
|
|
## 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. --> |
|
|
|
|
|
### Direct Use |
|
|
|
|
|
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> |
|
|
|
|
|
This model can be used directly for question-answering tasks, providing accurate and relevant answers based on the input queries. |
|
|
|
|
|
### Downstream Use [optional] |
|
|
|
|
|
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> |
|
|
|
|
|
The model can be fine-tuned for specific tasks or integrated into larger systems to enhance its capabilities and performance. |
|
|
|
|
|
### Out-of-Scope Use |
|
|
|
|
|
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> |
|
|
|
|
|
The model should not be used for generating harmful or biased content. It is not suitable for tasks requiring high levels of interpretability or transparency. |
|
|
|
|
|
## Bias, Risks, and Limitations |
|
|
|
|
|
<!-- This section is meant to convey both technical and sociotechnical limitations. --> |
|
|
|
|
|
The model may exhibit biases present in the training data. Users should be aware of these biases and take appropriate measures to mitigate them. |
|
|
|
|
|
### 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 is needed for further recommendations. |
|
|
|
|
|
## How to Get Started with the Model |
|
|
|
|
|
Use the code below to get started with the model. |
|
|
|
|
|
```python |
|
|
import os |
|
|
import openai |
|
|
|
|
|
# Set up OpenAI API key |
|
|
openai.api_key = os.getenv("OPENAI_API_KEY") |
|
|
|
|
|
# Generate a response |
|
|
response = openai.ChatCompletion.create( |
|
|
model="deepseek-ai/DeepSeek-V3", |
|
|
messages=[ |
|
|
{"role": "user", "content": "Your question here"} |
|
|
] |
|
|
) |
|
|
|
|
|
print(response.choices.message['content']) |
|
|
``` |
|
|
|
|
|
## Training Details |
|
|
|
|
|
### Training Data |
|
|
|
|
|
<!-- 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. --> |
|
|
|
|
|
The model has been trained on datasets such as DAMO-NLP-SG/multimodal_textbook, cognitivecomputations/dolphin-r1, open-thoughts/OpenThoughts-114k, PJMixers-Dev/open-thoughts_OpenThoughts-114k-CustomShareGPT, HumanLLMs/Human-Like-DPO-Dataset, Triangle104/HumanLLMs_Human-Like-DPO-Dataset, and fka/awesome-chatgpt-prompts. |
|
|
|
|
|
### Training Procedure |
|
|
|
|
|
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> |
|
|
|
|
|
The training procedure involved fine-tuning the base models using the provided datasets to enhance the model's performance in question-answering tasks. |
|
|
|
|
|
#### Preprocessing [optional] |
|
|
|
|
|
The data was preprocessed to ensure consistency and quality. This included tokenization, normalization, and filtering of irrelevant or noisy data. |
|
|
|
|
|
#### Training Hyperparameters |
|
|
|
|
|
- **Training regime:** fp16 mixed precision |
|
|
|
|
|
#### Speeds, Sizes, Times [optional] |
|
|
|
|
|
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> |
|
|
|
|
|
Training was conducted over a period of 72 hours using a cluster of NVIDIA A100 GPUs. The model checkpoints were saved every 12 hours. |
|
|
|
|
|
## Evaluation |
|
|
|
|
|
<!-- This section describes the evaluation protocols and provides the results. --> |
|
|
|
|
|
### Testing Data, Factors & Metrics |
|
|
|
|
|
#### Testing Data |
|
|
|
|
|
<!-- This should link to a Dataset Card if possible. --> |
|
|
|
|
|
The model was tested on a diverse set of question-answering benchmarks to evaluate its performance across different domains and query types. |
|
|
|
|
|
#### Factors |
|
|
|
|
|
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> |
|
|
|
|
|
The evaluation considered factors such as query complexity, domain specificity, and linguistic variations. |
|
|
|
|
|
#### Metrics |
|
|
|
|
|
<!-- These are the evaluation metrics being used, ideally with a description of why. --> |
|
|
|
|
|
The model has been evaluated using metrics such as character, accuracy, bertscore, code_eval, brier_score, bleu, and bleurt. |
|
|
|
|
|
### Results |
|
|
|
|
|
The model achieved high accuracy and robust performance across various benchmarks, demonstrating its effectiveness in question-answering tasks. |
|
|
|
|
|
#### Summary |
|
|
|
|
|
The model's performance metrics indicate strong capabilities in understanding and generating accurate responses to a wide range of queries. |
|
|
|
|
|
## Model Examination [optional] |
|
|
|
|
|
<!-- Relevant interpretability work for the model goes here --> |
|
|
|
|
|
The model's interpretability was assessed through attention visualization and feature importance analysis, providing insights into its decision-making process. |
|
|
|
|
|
## 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 *An external link was removed to protect your privacy.* presented in *An external link was removed to protect your privacy.*. |
|
|
|
|
|
- **Hardware Type:** NVIDIA A100 GPUs |
|
|
- **Hours used:** 72 hours |
|
|
- **Cloud Provider:** Azure |
|
|
- **Compute Region:** East US |
|
|
- **Carbon Emitted:** [More Information Needed] |
|
|
|
|
|
## Technical Specifications [optional] |
|
|
|
|
|
### Model Architecture and Objective |
|
|
|
|
|
The model is based on the transformer architecture and is designed to excel in question-answering tasks by leveraging large-scale pretraining and fine-tuning. |
|
|
|
|
|
### Compute Infrastructure |
|
|
|
|
|
The training and evaluation were conducted on a high-performance computing cluster with NVIDIA A100 GPUs. |
|
|
|
|
|
#### Hardware |
|
|
|
|
|
NVIDIA A100 GPUs |
|
|
|
|
|
#### Software |
|
|
|
|
|
The model was developed using Python, TensorFlow, and PyTorch frameworks. |
|
|
|
|
|
## 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:** |
|
|
|
|
|
```bibtex |
|
|
@misc{harrison2025deepseek, |
|
|
author = {Jonathan Harrison}, |
|
|
title = {DeepSeek: A Comprehensive Question-Answering Model}, |
|
|
year = {2025}, |
|
|
howpublished = {\url{https://github.com/deepseek-ai/DeepSeek-V3}}, |
|
|
} |
|
|
``` |
|
|
|
|
|
**APA:** |
|
|
|
|
|
Harrison, J. (2025). DeepSeek: A Comprehensive Question-Answering Model. Retrieved from https://github.com/deepseek-ai/DeepSeek-V3 |
|
|
|
|
|
## Glossary [optional] |
|
|
|
|
|
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> |
|
|
|
|
|
- **Transformer:** A type of neural network architecture that uses self-attention mechanisms to process input data. |
|
|
- **Fine-Tuning:** The process of further training a pre-trained model on a specific task or dataset to improve its performance. |
|
|
- **BERTScore:** A metric for evaluating the quality of text generation by comparing the similarity of embeddings between the generated text and reference text. |
|
|
|
|
|
## More Information [optional] |
|
|
|
|
|
For more details, visit the model's repository and documentation. |
|
|
|
|
|
## Model Card Authors [optional] |
|
|
|
|
|
[Jonathan Harrison] |
|
|
|
|
|
## Model Card Contact |
|
|
|
|
|
For inquiries, contact [Jonathan Harrison] at [email protected]. |
|
|
|
|
|
--- |
