list of pre-implemented models?
I see there are examples that run Llama2, Llama3, TinyLlama, and Stable Diffusion. Is there a list of pre-implemented models that can be used? Or is it just these examples? https://github.com/modularml/max/tree/main/examples/graph-api/pipelines/quantize_tinystories
GitHub
max/examples/graph-api/pipelines/quantize_tinystories at main · mod...
A collection of sample programs, notebooks, and tools which highlight the power of the MAX Platform - modularml/max
1 Reply
There are a few different ways to define a model for inference via MAX: in TorchScript, in ONNX, or construct in Mojo via the Graph API. We show several examples of TorchScript and ONNX models here: https://github.com/modularml/max/tree/main/examples/inference , which currently include BERT, Mistral 7B, ResNet-50, Stable Diffusion, and YOLOv8.
New in 24.4 are end-to-end pipelines that we've defined in Mojo and that use the MAX Graph API to construct the computational graph: https://github.com/modularml/max/tree/main/examples/graph-api/pipelines . We're referring to them as pipelines because the idea is that you can define all pre- and post-processing in Mojo as well (such as the tokenizer used in Llama 3) and easily incorporate them into a larger Mojo application. We've seeded this group with a few representative pipelines, and Llama 3 is the lead example among those.
We're extremely interested in having the community build upon these, as well as hearing what you'd like to see as additional examples, so please let us know how we can make this a better resource. We plan to regularly expand these examples.
GitHub
max/examples/inference at main · modularml/max
A collection of sample programs, notebooks, and tools which highlight the power of the MAX Platform - modularml/max
GitHub
max/examples/graph-api/pipelines at main · modularml/max
A collection of sample programs, notebooks, and tools which highlight the power of the MAX Platform - modularml/max