https://github.com/brevdev/notebooks
Collection of notebook guides created by the Brev.dev team!
https://github.com/brevdev/notebooks
Last synced: about 1 year ago
JSON representation
Collection of notebook guides created by the Brev.dev team!
- Host: GitHub
- URL: https://github.com/brevdev/notebooks
- Owner: brevdev
- License: mit
- Created: 2023-08-14T19:13:49.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2024-10-23T23:56:34.000Z (over 1 year ago)
- Last Synced: 2024-10-29T15:34:34.918Z (over 1 year ago)
- Language: Jupyter Notebook
- Homepage:
- Size: 79 MB
- Stars: 1,654
- Watchers: 26
- Forks: 284
- Open Issues: 9
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
Console •
Docs •
Templates •
Discord
# Brev.dev Notebooks
This repo contains helpful AI/ML notebook templates for LLMs, multi-modal models, image segmentation, and more. Each notebook has been coupled with the minimum GPU specs required to use them along with a 1-click deploy badge that starts each notebook on a GPU.
## Contributing
We welcome contributions to this repository! If you have a notebook you'd like to add, please reach out to use the [Discord](https://discord.gg/y9428NwTh3) or open a pull request!
## Notebooks
We've split the notebooks into categories based on the type of task they're designed for. Our current split is: LLM finetuning/training, multi-modal models, computer vision/image segmentation, and miscellaneous. Let us know if you want to see more notebooks for a certain task or using different frameworks and tools!
### LLM Finetuning/Training
| Notebook | Description | Min. GPU | Deploy |
| ------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Fine-tune Llama3 using Direct Preference Optimization](https://github.com/brevdev/notebooks/blob/main/llama3dpo.ipynb) | Fine-tune Llama3 using DPO | 1x A100 | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/llama3dpo.ipynb) [](https://console.brev.dev/notebook/llama3dpo) |
| [Fine-tune Llama3 using SFT](https://github.com/brevdev/notebooks/blob/main/llama3_finetune_inference.ipynb) | Fine-tune and deploy Llama 3 | 2x A100 | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/llama3_finetune_inference.ipynb) [](https://console.brev.dev/notebook/llama3_finetune_inference) |
| [Fine-tune Llama 2](https://github.com/brevdev/notebooks/blob/main/llama2-finetune.ipynb) | A Guide to Fine-tuning Llama 2 | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/llama2-finetune.ipynb) [](https://console.brev.dev/notebook/llama2-finetune) [](https://www.youtube.com/watch?v=lPLrODJjHUE) |
| [Fine-tune Llama 2 - Own Data](https://github.com/brevdev/notebooks/blob/main/llama2-finetune-own-data.ipynb) | Fine-tune Llama 2 on your own dataset | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/llama2-finetune-own-data.ipynb) [](https://console.brev.dev/notebook/llama2-finetune-own-data) [](https://www.youtube.com/watch?v=lPLrODJjHUE) |
| [Fine-tune Mistral](https://github.com/brevdev/notebooks/blob/main/mistral-finetune.ipynb) | A Guide to Fine-tuning Mistral | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/mistral-finetune.ipynb) [](https://console.brev.dev/notebook/mistral-finetune) |
| [Fine-tune Mistral using NVIDIA NeMO and PEFT](https://github.com/brevdev/notebooks/blob/main/mistral-finetune-nemo.ipynb) | Fine-tune Mistral using NVIDIA NeMO and PEFT | 1x A100 | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/mistral-finetune-nemo.ipynb) [](https://console.brev.dev/notebook/mistral_nemo_finetune) |
| [Fine-tune Mistral - Own Data](https://github.com/brevdev/notebooks/blob/main/mistral-finetune-own-data.ipynb) | Fine-tune Mistral on your own dataset | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/mistral-finetune-own-data.ipynb) [](https://console.brev.dev/notebook/mistral-finetune-own-data) [](https://www.youtube.com/watch?v=kmkcNVvEz-k) |
| [Fine-tune Mixtral (8x7B MoE)](https://github.com/brevdev/notebooks/blob/main/mixtral-finetune.ipynb) | A Guide to Fine-tuning Mixtral, Mistral's 8x7B MoE | 4x T4 | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/mixtral-finetune.ipynb) [](https://console.brev.dev/notebook/mixtral-finetune-own-data) [](https://www.youtube.com/watch?v=zbKz4g100SQ) |
| [Fine-tune Mixtral (8x7B MoE) - Own Data](https://github.com/brevdev/notebooks/blob/main/mixtral-finetune-own-data.ipynb) | A Guide to Fine-tuning Mixtral on your own dataset | 4x T4 | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/mixtral-finetune-own-data.ipynb) [](https://console.brev.dev/notebook/mixtral-finetune-own-data) |
| [Fine-tune BioMistral](https://github.com/brevdev/notebooks/blob/main/biomistral-finetune.ipynb) | A Guide to Fine-tuning BioMistral | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/biomistral-finetune.ipynb) [](https://console.brev.dev/environment/new?instance=A10G:g5.xlarge&name=biomistral-finetune&file=https://github.com/brevdev/notebooks/raw/main/biomistral-finetune.ipynb&python=3.10&cuda=12.0.1) |
| [Fine-tune Phi-2](https://github.com/brevdev/notebooks/blob/main/phi2-finetune.ipynb) | A Guide to Fine-tuning Phi-2 | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/phi2-finetune.ipynb) [](https://console.brev.dev/notebook/phi2-finetune-own-data) [](https://www.youtube.com/watch?v=t55XrJddjLA) |
| [Fine-tune Phi-2 - Own Data](https://github.com/brevdev/notebooks/blob/main/phi2-finetune-own-data.ipynb) | Fine-tune Phi-2 on your own dataset | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/phi2-finetune-own-data.ipynb) [](https://console.brev.dev/notebook/phi2-finetune-own-data) [](https://www.youtube.com/watch?v=t55XrJddjLA) |
| [Training Question/Answer models using NVIDIA NeMo](https://github.com/brevdev/notebooks/blob/main/question_answer_nemo.ipynb) | Use NeMo to train BERT, GPT, and S2S models for Q&A tasks | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/question_answer_nemo.ipynb) [](https://console.brev.dev/notebook/question_answer_nemo) |
### LLM Inference/Deployment
| Notebook | Description | Min. GPU | Deploy |
| ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| [Run inference on Llama3 using TensorRT-LLM](https://github.com/brevdev/notebooks/blob/main/tensorrt-llama3.ipynb) | Run inference on Llama3 using TensorRT-LLM | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/tensorrt-llama3.ipynb) [](https://console.brev.dev/notebook/llama3-tensorrtllm-deployment) |
| [Inference on DBRX with VLLM and Gradio](https://github.com/brevdev/notebooks/blob/main/dbrx_inference.ipynb) | Run inference on DBRX with VLLM and Gradio | 4x A100 | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/deploy-to-replicate.ipynb) [](https://console.brev.dev/notebooks/dbrx_inference) |
| [Run BioMistral](https://github.com/brevdev/notebooks/blob/main/biomistral.ipynb) | Run BioMistral | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/biomistral.ipynb) [](https://console.brev.dev/environment/new?instance=A10G:g5.xlarge&name=biomistral&file=https://github.com/brevdev/notebooks/raw/main/biomistral.ipynb&python=3.10&cuda=12.0.1) |
| [Run Llama 2 70B](https://github.com/brevdev/notebooks/blob/main/llama2-finetune-own-data.ipynb) | Run Llama 2 70B, or any Llama 2 Model | 4x T4 | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/llama2.ipynb) [](https://console.brev.dev/notebooks/llama2-finetune-own-data) |
| [Use TensorRT-LLM with Mistral](https://github.com/brevdev/notebooks/blob/main/tensorrt_mistral.ipynb) | Use NVIDIA TensorRT engine to run inference on Mistral-7B | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/tensorrt_mistral.ipynb) [](https://console.brev.dev/notebook/trtmistral1) |
| [StreamingLLM for Optimized Inference](https://github.com/brevdev/notebooks/blob/main/streamingllm-tensorrt.ipynb) | Use StreamingLLM for infinite length input without finetuning | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/streamingllm-tensorrt.ipynb) [](https://console.brev.dev/notebook/streamingllm-tensorrt-llm) |
### Multi-modal and Computer Vision Models
| Notebook | Description | Min. GPU | Deploy |
| ------------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------- | -------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Finetune and deploy LlaVA](https://github.com/brevdev/notebooks/blob/main/llava-finetune.ipynb) | Finetune the LlaVA model on your own data | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/llava-finetune.ipynb) [](https://console.brev.dev/notebook/llava-finetune) [](https://www.youtube.com/watch?v=XICHJx2_Rm8) |
| [AUTOMATIC1111 Stable Diffusion WebUI](https://github.com/brevdev/notebooks/blob/main/automatic1111-stable-diffusion-ui.ipynb) | Run Stable Diffusion WebUI, AUTOMATIC1111 | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/stable-diffusion-ui.ipynb) [](https://console.brev.dev/notebook/automatic1111-stable-diffusion-ui) [](https://www.youtube.com/watch?v=Sf6PwCz6fbI) |
| [ControlNet on AUTOMATIC1111](https://github.com/brevdev/notebooks/blob/main/controlnet.ipynb) | Run ControlNet Models on Stable Diffusion WebUI | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/controlnet.ipynb) [](https://console.brev.dev/notebook/controlnet) |
| [SDXL inference with LoRA and Diffusers](https://github.com/brevdev/notebooks/blob/main/diffusion_lora_inference.ipynb) | Run inference using LoRA adaptors and SDXL | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/diffusion_lora_inference.ipynb) [](https://console.brev.dev/notebook/diffusion_lora_inf) |
| [Oobabooga LLM WebUI](https://github.com/brevdev/notebooks/blob/main/oobabooga.ipynb) | Run Oobabooga, the LLM WebUI (like AUTOMATIC1111) | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/oobabooga.ipynb) [](https://console.brev.dev/notebook/oobabooga) |
| [EfficientViT Segement Anything](https://github.com/brevdev/notebooks/blob/main/efficientvit-segmentation.ipynb) | Run a TensorRT optimized version of Segment Anything | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/efficientvit-segmentation.ipynb) [](https://console.brev.dev/notebook/efficientvit-segmentation) |
### Other applications and tools
| Notebook | Description | Min. GPU | Deploy |
| ---------------------------------------------------------------------------------------------------------------------- | ------------------------------------------- | ------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Deploy to Replicate](https://github.com/brevdev/notebooks/blob/main/deploy-to-replicate.ipynb) | Deploy Model to Replicate | any \|\| CPU | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/deploy-to-replicate.ipynb) [](https://console.brev.dev/notebook/deploy-to-replicate) [](https://www.youtube.com/watch?v=eczHFcqx1ic) |
| [GGUF Export FT Model](https://github.com/brevdev/notebooks/blob/main/gguf-export.ipynb) | Export your fine-tuned model to GGUF | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/gguf-export.ipynb) [](https://console.brev.dev/notebook/gguf-export) |
| [Julia Install](https://github.com/brevdev/notebooks/blob/main/julia-install.ipynb) | Easily Install Julia + Notebooks | any \|\| CPU | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/julia-install.ipynb) [](https://console.brev.dev/notebook/julia-install) |
| [PDF Chatbot (OCR)](https://github.com/brevdev/notebooks/blob/main/ocr-pdf-analysis.ipynb) | PDF Chatbot using OCR | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/ocr-pdf-analysis.ipynb) [](https://console.brev.dev/notebooks/ocr-pdf-analysis) |
| [Zephyr Chatbot](https://github.com/brevdev/notebooks/blob/main/zephyr-chatbot.ipynb) | Chatbot with Open Source Models | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/zephyr-chatbot.ipynb) [](https://console.brev.dev/notebook/zephyr-chatbot) |
| [Accelerate Data Science using NVIDIA RAPIDS](https://github.com/brevdev/notebooks/blob/main/rapids_cudf_pandas.ipynb) | Accelerate Data Science using NVIDIA RAPIDS | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/rapids_cudf_pandas.ipynb) [](https://console.brev.dev/notebook/rapids_cudf_pandas) |
| [Accelerate Data Science using NVIDIA RAPIDS](https://github.com/brevdev/notebooks/blob/main/rapids_cudf_pandas.ipynb) | Accelerate Data Science using NVIDIA RAPIDS | 1x A10G | [](https://colab.research.google.com/github/brevdev/notebooks/blob/main/rapids_cudf_pandas.ipynb) [](https://console.brev.dev/notebook/rapids_cudf_pandas) [](https://console.brev.dev/notebook/rapids_cudf_pandas) |
---
### What is Brev.dev?
Brev is a dev tool that makes it really easy to code on a GPU in the cloud. Brev does 3 things: provision, configure, and connect.
#### Provision:
Brev provisions a GPU for you. You don't have to worry about setting up a cloud account. We have solid GPU supply, but if you do have AWS or GCP, you can link them.
#### Configure:
Brev configures your GPU with the right drivers and libraries. Use our open source tool Verb to point and click the right python and CUDA versions.
#### Connect:
Brev.dev CLI automatically edits your ssh config so you can `ssh gpu-name` or run `brev open gpu-name` to open VS Code to the remote machine