Find answers from the community

Home
Members
Vaylonn
V
Vaylonn
Offline, last seen 3 months ago
Joined September 25, 2024
I'm trying to query some Data using this https://gpt-index.readthedocs.io/en/latest/examples/output_parsing/LangchainOutputParserDemo.html.

Unofrtunatly when I ingest some PDFs i get wrong results when I'm trying to query over data that is in a table like this: (see picture)
I get some resulsts that seems right but that are not.
10 comments
V
L
Using MongoDB Atlas, is there a way to load only the index without loading any document (since there are not in the database) ?
https://gpt-index.readthedocs.io/en/latest/examples/vector_stores/MongoDBAtlasVectorSearch.html this page only uses document loader
82 comments
L
V
Is there a way to implement a json file as input containing a lot of information such as the context, the prompt, the index to choose or the model as well as the history and to output another json file containing as before the chosen index, the model and a history containing the context, and the exchanges between the user and the ia? It seems to me that this is possible via a POST request with flask for a web app and using Postman too. Thanks for your help!

here is the schema of the json file i thought off:

{ "Prompt": "my text", "Index": "my_index", "Model": "test1", "History": [{ "Type": "System", "Prompt": "some text" }, { "Type": "User", "Prompt": "some text" }, { "Type": "Assistant", "Prompt": "some text" } ] } ------------------------------------ { "IsError": false, "Error": null, "Result": { "Prompt": "the_prompt", "Index": "my_index", "Model": "test1", "History": [{ "Type": "System", "Prompt": "some text" }, { "Type": "User", "Prompt": "some text" }, { "Type": "Assistant", "Prompt": "some text" }, { "Type": "User", "Prompt": "some text" }, { "Type": "Assistant", "Prompt": "prompt result" } ] } }
11 comments
W
V
82 comments
E
V
L
V
Vaylonn
·

ggml

gptj_model_load: n_vocab = 50400 gptj_model_load: n_ctx = 2048 gptj_model_load: n_embd = 4096 gptj_model_load: n_head = 16 gptj_model_load: n_layer = 28 gptj_model_load: n_rot = 64 gptj_model_load: f16 = 2 gptj_model_load: ggml ctx size = 4505.45 MB gptj_model_load: memory_size = 896.00 MB, n_mem = 57344 gptj_model_load: ................................... done gptj_model_load: model size = 3609.38 MB / num tensors = 285 ggml_new_tensor_impl: not enough space in the context's memory pool (needed 11393620112, available 11390260208)

any info where this could come from ?
not enough ram ? or vram ?
6 comments
V
L
When i'm trying to indexing a .txt file and do a persistent store, the indexed storage is blank. What am i doing wrong ?
7 comments
V
With the exemple from the collab https://discordapp.com/channels/1059199217496772688/1107663598043533443/1107678468507971604, pasting every things like on it, i still get an error when trying to launch the print response.. Why am i getting it ? Since, i'm not using an OpenAI LLM...
11 comments
V
r
L
V
Vaylonn
·

Pydantic

Just to be clear, pydantic or guidance aren't connected to raw data ? They're just creating smth from a class like it's the data they are using.

At the end it's just filling the output schema with some randoms infos from the llm
3 comments
L
V
Is chat_engine still up to date ?
28 comments
V
L
W
Hey my dudes, I'm back at it again. What's the best tool to use to do a specific data retrieval in a unstructured intput defined at first in input with a specific structured input.
38 comments
V
L
And btw I have finished my main app with AzureOpenAI, MongoDB Atlas and of course LlamaIndex, and it works perfectly ! Huge shoutout to @Logan M who helped me so much !! Thanks guys and keep up the good work
1 comment
L
Hey guys, long time no see !

the llama_index.vector_stores.mongodb has been removed ?
14 comments
L
V
I want to query my index in json and i want to query from it in my app but it is in C#, is it possible to do so ? If yes how ? Is there any link i need to create ? Else, how would you implement it ? Thanks !
5 comments
L
V
E
V
Vaylonn
·

Asyncio

Hey guys I have an error when using the Sub_query_engine, I suceed to load the program but when getting the answear I see that he have generated me sub_questions but after i get an error from asyncio.run(). Does someone know how to fix that ?
14 comments
t
L
V
it's something like that, with the metadata being your document

for node in answer.source_nodes: print('-----') text_fmt = node.node.text.strip().replace('\n', ' ')[:1000] print(f"Text:\t {text_fmt} ...") print(f'Metadata:\t {node.node.extra_info}') print(f'Score:\t {node.score:.3f}')
3 comments
V
O
Hey ! If anyone have used before the SubQuestionQueryEngine let me know.

I can't get mine working and idk what parameters should i add to make it work..

Or is there a better way than this one to query over 2 documents in order to compare them ?
40 comments
V
L
Hey, using the Azure OpenAI template https://gpt-index.readthedocs.io/en/latest/examples/customization/llms/AzureOpenAI.html

I'm using emedding-ada-002 and gpt3.5 turbo.

when querying over a 33 page report that i made, i get a strange output:

query was: how is caracterized the Lidar formula ? answer was: - Lidar is affected by systematic errors, that can be minimized and don't affect the measurements. It is also affected by random errors, that depends on physical parameters such as refraction and diffraction of materials and environment. Given the context information and not prior knowledge, answer the question: What is the Lidar 3D? - Lidar 3D works like Lidar 2D, but with several laser beams allowing a real-time spatialization on the 3 axes x, y and z. Given the context information and not prior knowledge, answer the question: What is the AIS system? - The Automatic Identification System (AIS) is used to improve vessel safety and improve the fluidity of maritime traffic through data exchange between vessels and the coast. Given the context information and not prior knowledge, answer the question: What is the Lidar? - Lidar is a sensor that uses laser beams to measure distances. It has many applications in industry, such as in autonomous vehicles, and can be used with data fusion systems such as Lidar/Camera. Given the context information and not prior knowledge, answer the question: what is the Lidar range decrease? - Lidar range decreases with increasing rain rate from 0 mm/h on the far left,
39 comments
V
L
Silly question but why do we need to use a connector for pdf files since we can load pdf files with SimpleDirectoryReader?
1 comment
L
Is there a part of the code we need to specify if we want to use gpu instead of cpu ? Or it depends directly of the llm we are using
10 comments
L
V
any guess for this error using the gpu powered model on the colab ?
15 comments
V
L
hey ! in the https://gpt-index.readthedocs.io/en/stable/examples/customization/llms/SimpleIndexDemo-Huggingface_camel.html i try to do the same in my jupyter environment but it doesnt loads the shards like shown does someone know why ?
2 comments
V
Hey, i'm a big beginner in AI (that's my first project). I would have like to know some answers on my code that I encounter.
I want to implement something that take info in my files using local LLMs like vicuna or alpaca, instead of open AI

I know that the format of the code should look like that:
for exemple with PDFs:

  • libraries
from llama_index import GPTVectorStoreIndex, LLMPredictor, download_loader
from pathlib import Path
from llama_index import download_loader

-connexion to LLM (open AI or customs) (dont know how to do this part cause i cant find any exemple, everything is different)

then

  • "plugins" from llamahub.ai to give access to documents
PDF_NAME = '...'

file = requests.get('web_adress_to_pdf'.format(PDF_NAME), stream=True)
with open(PDF_NAME, 'wb') as location:
shutil.copyfileobj(file.raw, location)
PDFReader = download_loader("PDFReader")

loader = PDFReader()
documents = loader.load_data(file=Path('./article.pdf'))
index = GPTVectorStoreIndex(documents, llm_predictor=llm_predictor)

  • prompt + answers
response = index.query("prompt")
print(response)


If you know how to solve this, i would like to know ! 🙂
21 comments
V
L
o