10.Expand LLaVA-Phi3
10.Expand LLaVA-Phi3Model scalePerformanceGot LLaVA-Phi3Use LLaVA-Phi3Run LLaVA-Phi3Have a conversationEnd the conversationReferences
Demo Environment
Development Boards：Jetson nano
SD(TF) card：64G
xIt is recommended to run the 4B and below parameter models
LLaVA-Phi3 is a LLaVA model fine-tuned from the Phi 3 Mini 4k.
xLLaVA (Large-scale Language and Vision Assistant) is a multimodal model that aims to achieve general vision and language understanding by combining a visual encoder and a large-scale language model.
Model scale
| Model | Parameter |
|---|---|
| LLaVA-Phi3 | 3.8B |
Performance
Got LLaVA-Phi3
Using the pull command will automatically pull the model from the Ollama model library.
xxxxxxxxxxollama pull llava-phi3:3.8b
Use LLaVA-Phi3
Use LLaVA-Phi3 to identify local image content.
Run LLaVA-Phi3
If the system does not have a running model, the system will automatically pull the LLaVA-Phi3 3.8B model and run it.
xxxxxxxxxxollama run llava-phi3:3.8b
If an error occurs during operation, you can restart the system and try again.
Have a conversation
xxxxxxxxxxWhat's in this image? /home/jetson/Pictures/1.jpeg
The response time is related to the hardware configuration, please be patient.
End the conversation
Use the Ctrl+d shortcut or /bye to end the conversation.
References
Ollama
Website：https://ollama.com/
GitHub：https://github.com/ollama/ollama
LLaVA-Phi3
GitHub：https://github.com/InternLM/xtuner/tree/main
Ollama corresponding model：https://ollama.com/library/llava-phi3