It just got a lot easier to get started with Groq technology with GroqFlow! GroqFlow, our automatic toolflow for mapping machine learning workloads to GroqChip™, is now launched on GitHub.
Watch this 3-minute video to learn how you can get started with just one line of code, then head over to GitHub to check it out for yourself.
If you didn’t catch our GroqFlow announcement, you can read it here.
Below is a snapshot of our fast-growing list of models that we’ve already tested with GroqFlow with out-of-the-box support and no modifications needed. If you’re a Groq customer, download GroqFlow today.
Graph Convolutions from PyTorchGeometric:
- GINEconv
- GMMConv
- PNAConv
- SAGEConv
- GINConv
- CGConv
Huggingface
Transformers:
- BERT
- CamemBERT
- ConvNeXt
- DeBERTa
- DistilBERT
- Electra
- GPT-1
- GPT-2
- I-BERT
- LayoutLM
- LUKE
- MiniLMv2
- MobileBERT
- REALM
- RoBERTa
- SegFormer
- Speech2Text
- Splinter
- T5 Encoder
- XLM-RoBERTa
TorchVision Image Classification
- AlexNet
- GhostNet
- GoogLeNet
- HarDNet 39-DS/68/68-DS
- MEAL V1/V2
- MobileNet v2
- ProxylessNAS CPU/GPU/Mobile
- ResNet 18/34/50/101
- Shufflenet V2 x1.0
- SqueezeNet 1.0/1.1