whatsapp-chatgpt
whatsapp-chatgpt copied to clipboard
askrella
Latest gpt-4-1106-preview addition to the repo
Would it be possible to add it to the script? It seems to be pretty faster then GPT-4 and 2x - 3x cheaper too. Also seems to be updated to have knowledge up to April 2023 as compared to 2021. Also Dall-E v3 is also available now, which WAY better then Dall-E v2.
-
goto .env and udpate your "OPENAI_GPT_MODEL=gpt-4-vision-preview"
-
goto \whatsapp-chatgpt-master\src\handlers\dalle.ts... line 28 onwards replace the response: const response = await openai.createImage({ model: "dall-e-3", prompt: prompt, n: 1, //size: aiConfig.dalle.size as CreateImageRequestSizeEnum, size: "1024x1024", response_format: "b64_json" });
goto .env and udpate your "OPENAI_GPT_MODEL=gpt-4-vision-preview"
- goto \whatsapp-chatgpt-master\src\handlers\dalle.ts... line 28 onwards replace the response: const response = await openai.createImage({ model: "dall-e-3", prompt: prompt, n: 1, //size: aiConfig.dalle.size as CreateImageRequestSizeEnum, size: "1024x1024", response_format: "b64_json" });
Thank you for that. DALL-E 3 is working fine, but the GPT-Model still seems to be GPT 3.5 or at least it refused to answer anything after September 2021 date by stating
My "knowledge" is based on data up to my last update in September 2021, and I generate responses based on patterns and information from that dataset.
Same here, were you able to fix it or at least check that its actually using the right model?
Same here, were you able to fix it or at least check that its actually using the right model?
No, still stuck on 3.5.
Same here, were you able to fix it or at least check that its actually using the right model?
No, still stuck on 3.5.
So managed to get Dalle-3 to work but my gpt is still saying it's cutoff is 2021 even though the usage page shows that it's actually using the new model.