Hey guys. So glad you decided to post. Hard to believe I am first viewer! Subscribed :-) Also; Harrison = LangChain = WOW !!! The question at the ~45 minute mark regarding demand for local NON Public API LLMs that avoid exfiltration. Yes, must have for enterprise solutions, especially when automating back office. Great session.
Great video! I really appreciate the support for the growth and progress in the current situation. It's interesting to see that Yasser is the one sharing firsthand experience with implementing QA answering using langchain in a real-world business context. Looking forward to hearing from other attendees in future webinars. Shoutout to @Harrison for asking thought-provoking questions that drive the community forward!
Good video . I disagree comparing sending data to API compares to send it to AWS for building your e-commerce. Building your e-commerce on AWS they don’t care about your data, Amazon wants you to consume their infra. OpenAI or Anthropic need your data, those models feed from data, that is a fundamental difference on privacy to care about
I am building a chat with documents application using Langchain i am specially using ConversationalRetrievalQAChain with returnSourceDocuments set to true, but when i ask general questions like "Who is the CEO of twitter?" the answer is coming from llm(openai), with the source document, where this information is not mentioned. Also for greetings like "hi", "hello", etc. the source document is fetched, i only want to show the source document when the answer is present in it. How can i achieve the demo shown at 33:12 ? Like for greetings i don't want to generate/show source documents ?
Hi first of all great Webinar! If a Question is out of Context and the Information is not in the Document, the Chain rightfully says "I can't answer the Question", but is there a way for the Chain to try to answer the Question using its own language model, for questions outside of the context in the given Documents ?
Question (for anybody please!). With regards to PDF question answering with langchain. Can only completion openai models be used at the moment? or can chat as well? As get the double input error: /v1/chat/completions gpt-4, gpt-4-0314, gpt-4-32k, gpt-4-32k-0314, gpt-3.5-turbo, gpt-3.5-turbo-0301 /v1/completions text-davinci-003, text-davinci-002, text-curie-001, text-babbage-001, text-ada-001. Would love to use gpt-3.5-turbo as it's quite a bit cheaper while still being on par with davinci. If anybody has any insight please let me know. Thank you!
This is because you are using the wrong model, probably or the wrong call. Check your code and see if you are using a chat model for completion. Also, verify openai official documentation. I wish that helps. 😊
Hey guys. So glad you decided to post. Hard to believe I am first viewer! Subscribed :-)
Also; Harrison = LangChain = WOW !!!
The question at the ~45 minute mark regarding demand for local NON Public API LLMs that avoid exfiltration. Yes, must have for enterprise solutions, especially when automating back office.
Great session.
Great video! I really appreciate the support for the growth and progress in the current situation. It's interesting to see that Yasser is the one sharing firsthand experience with implementing QA answering using langchain in a real-world business context. Looking forward to hearing from other attendees in future webinars. Shoutout to @Harrison for asking thought-provoking questions that drive the community forward!
Good video . I disagree comparing sending data to API compares to send it to AWS for building your e-commerce. Building your e-commerce on AWS they don’t care about your data, Amazon wants you to consume their infra. OpenAI or Anthropic need your data, those models feed from data, that is a fundamental difference on privacy to care about
Tbh this was so much better than i thought it would be!😂
Great work guys, super valuable info. Thanks!
Great, thank you for a webinar
Great conversation!, thanks!
I am building a chat with documents application using Langchain i am specially using ConversationalRetrievalQAChain with returnSourceDocuments set to true, but when i ask general questions like "Who is the CEO of twitter?" the answer is coming from llm(openai), with the source document, where this information is not mentioned. Also for greetings like "hi", "hello", etc. the source document is fetched, i only want to show the source document when the answer is present in it. How can i achieve the demo shown at 33:12 ? Like for greetings i don't want to generate/show source documents ?
Hi first of all great Webinar!
If a Question is out of Context and the Information is not in the Document, the Chain rightfully says "I can't answer the Question", but is there a way for the Chain to try to answer the Question using its own language model, for questions outside of the context in the given Documents ?
Question (for anybody please!). With regards to PDF question answering with langchain. Can only completion openai models be used at the moment? or can chat as well? As get the double input error: /v1/chat/completions gpt-4, gpt-4-0314, gpt-4-32k, gpt-4-32k-0314, gpt-3.5-turbo, gpt-3.5-turbo-0301
/v1/completions text-davinci-003, text-davinci-002, text-curie-001, text-babbage-001, text-ada-001. Would love to use gpt-3.5-turbo as it's quite a bit cheaper while still being on par with davinci. If anybody has any insight please let me know. Thank you!
This is because you are using the wrong model, probably or the wrong call. Check your code and see if you are using a chat model for completion. Also, verify openai official documentation. I wish that helps. 😊