MAX Builds

Tags

Featured

A combination of what’s new, what’s popular, and what we think you’ll like this week.

DeepSeek Coder

This pipeline provides optimized support for the DeepSeek Coder V1.5 code completion model, a member of the LlamaForCausalLM architecture family. DeepSeek Coder is an open source code generation model released by DeepSeek.

Llama 3.1

This pipeline provides optimized support for the LlamaForCausalLM family of large language models, as exemplified by the Llama 3.1 text completion model. The structure of this implementation of the model was inspired by Andrej Karpathy's llama.c, and originally written in Mojo by Aydyn Tairov. The model itself has been constructed from end to end in the Mojo language using the MAX Graph API.

Text Generation

INT4

Transformer

Tags

Model Repository

Models that have been optimized for the MAX platform to run super fast.

Llama 3.1

@Modular

This pipeline provides optimized support for the LlamaForCausalLM family of large language models, as exemplified by the Llama 3.1 text completion model. The structure of this implementation of the model was inspired by Andrej Karpathy's llama.c, and originally written in Mojo by Aydyn Tairov. The model itself has been constructed from end to end in the Mojo language using the MAX Graph API.

Replit

An open source code generation model trained on permissively licensed code and released by Replit. The V1.5, 3B variant is the basis for this implementation, and weights are obtained via Hugging Face. This pipeline demonstrates code completion from an initial prompt using Replit's Code V1.5 3B large language model.

Mistral

This pipeline provides optimized support for the MistralForCausalLM family of large language models, as exemplified by the Mistral NeMo 12B text completion model. Mistral NeMo is an open source large language model released by Mistral AI in partnership with NVIDIA.

DeepSeek Coder

This pipeline provides optimized support for the DeepSeek Coder V1.5 code completion model, a member of the LlamaForCausalLM architecture family. DeepSeek Coder is an open source code generation model released by DeepSeek.

Code Completion

Transformer

DeepSeek

Tags

Packages

Reusable projects built by the Modular community on MAX and Mojo.

Tags

All

All items in this site in alphabetical order.

DeepSeek Coder

@Modular

This pipeline provides optimized support for the DeepSeek Coder V1.5 code completion model, a member of the LlamaForCausalLM architecture family. DeepSeek Coder is an open source code generation model released by DeepSeek.

Llama 3.1

This pipeline provides optimized support for the LlamaForCausalLM family of large language models, as exemplified by the Llama 3.1 text completion model. The structure of this implementation of the model was inspired by Andrej Karpathy's llama.c, and originally written in Mojo by Aydyn Tairov. The model itself has been constructed from end to end in the Mojo language using the MAX Graph API.

Text Generation

INT4

Transformer

Build AI apps fast on MAX

Build AI apps fast on MAX

Featured

Model Repository

Packages

All

What is MAX Graph?

Our goal is to make MAX Graphs easy to use, so please give us feedback and help us improve this for the industry.

for new releases

for tutorials