Well done, fast and easy to follow explanation of a powerful and useful case. Another useful case is combing the content of more than 1 file and creating reports or explainers. Thank you
Thank you! I’m really glad you found it helpful. That’s a great suggestion! Combining multiple files and generating reports or explainers would make for an awesome follow-up tutorial. Stay tuned, I’ll look into covering that in a future video! Appreciate your support!
hello ! I have a question, I did exactly what you have done in the video but when i upload a file, the chatbot answer is "Hmm, I'm not sure." what can I do ?
Thanks for following along! The error could be due to the chunk size of the file. Try adjusting the chunk sizes based on the file size you're uploading-sometimes a bit of tweaking can make a big difference. Also, check the 'Top K' settings in the Pinecone node to see if adjusting that helps improve responses. Let me know if that helps.
@TheAIAgentGuy Hi. I have the same problem. I tried everything with chunk sizes and Top K, however it seems that this is not an issue. Any others ideas?
Indexes must be created with a dimension that matches the model’s embedding dimension. If you use the OpenAI embeddings model, you have to set the dimensions to 1536 in Pinecone. You cannot change the dimensions afterward, so you have to create a new index with the correct dimensions. Hope this helps.
Did you get it working with ollama? @MeTuMaTHiCa I am using nomic-embed-text as rhe ollama embedding and llava:7b as the chat model, but I'm getting errors... What did you use with ollama for it to work?
Well done, fast and easy to follow explanation of a powerful and useful case. Another useful case is combing the content of more than 1 file and creating reports or explainers. Thank you
Thank you! I’m really glad you found it helpful.
That’s a great suggestion! Combining multiple files and generating reports or explainers would make for an awesome follow-up tutorial. Stay tuned, I’ll look into covering that in a future video!
Appreciate your support!
That's great. What would be nice is to have a list of uploaded files available, so that you can add/delete at will.
Great job.
hello ! I have a question, I did exactly what you have done in the video but when i upload a file, the chatbot answer is "Hmm, I'm not sure." what can I do ?
Thanks for following along! The error could be due to the chunk size of the file. Try adjusting the chunk sizes based on the file size you're uploading-sometimes a bit of tweaking can make a big difference. Also, check the 'Top K' settings in the Pinecone node to see if adjusting that helps improve responses. Let me know if that helps.
@TheAIAgentGuy Hi. I have the same problem. I tried everything with chunk sizes and Top K, however it seems that this is not an issue. Any others ideas?
Hello, I have tried Pinecone yet I don't have the file upload yet on my flowise
Hi, you might need to update your Flowise version. Its a relatively new feature.
Can we upload doc or doc files?
Yes it's possible! Check this out: docs.flowiseai.com/using-flowise/uploads
my warning :Vector dimension 768 does not match the dimension of the index 1536, why does this happen
Indexes must be created with a dimension that matches the model’s embedding dimension.
If you use the OpenAI embeddings model, you have to set the dimensions to 1536 in Pinecone. You cannot change the dimensions afterward, so you have to create a new index with the correct dimensions.
Hope this helps.
@@TheAIAgentGuy thx for information and time u spent. i am using ollama embeddings, what may be the dimensions, or how can i find dimensions
@@TheAIAgentGuy ok thx for embedding warning, i found nomic embeddings dimension: 768.
thx a lot
Did you get it working with ollama? @MeTuMaTHiCa
I am using nomic-embed-text as rhe ollama embedding and llava:7b as the chat model, but I'm getting errors...
What did you use with ollama for it to work?
Error: vectorsService.upsertVector - Error: Error loading file
Any clue why that might be?
Hard to tell without taking a look at it. Check your index key for something missing, perhaps.