- Playground-like controls (continue, regen, reset)
- Optional additional file input (pdf, txt, html, etc.)
- Recognizes group conversations with usernames
- Image recognition with Q&A
- Markdown formatting (italic, bold, underline, emoji, code, quote, list)
- Codex with automatic code formatting
- Complex mathematical expression rendering
- Charts and diagrams generation
- Safe image search capabilities
- Two-Part Memory System:
- Short-Term Adaptive Memory (typical)
- Long-Term Adaptive Memory (like BlenderBot)
#Testing a GPT-3 Discord Bot with a Growing List of Features (Examples Attached)
15 messages · Page 1 of 1 (latest)
@regal pike
You have been muted by AutoMod.
more examples, including file input
Woah, awesome work! How do you get information from the photos? CLIP? Or something else?
How can i use this?
I have a similar bot not nearly as advanced as this. Tell me when its availible to add to my server. Its awesome!
thank you yes we use BLIP
you can join our support server to interact with Iva and/or add the bot to your own server
yes absolutely
Self hosted or with API? I want to use blip 2 :D
API. we would like to use BLIP2 too 😄
Can you please share if possible which API? I have used two different ones from replicate which are decent ish
const initialResult = await axios.post<ReplicatePrediction<TImageDecoder>>(`https:// api.replicate.com/v1/predictions`, {
version: 'a4a8bafd6089e1716b06057c42b19378250d008b80fe87caa5cd36d40c1eda90',
input: {
image: url,
clip_model_name: 'ViT-H-14/laion2b_s32b_b79k',
mode: 'fast',
}
}, {
headers: {
'Authorization': `Token ${REPLICATE_API_KEY}`,
'Content-Type': 'application/json'
}
});
we use replicate!