h2ogpt
h2ogpt copied to clipboard
Private chat with local GPT with document, images, video, etc. 100% private, Apache 2.0. Supports oLLaMa, Mixtral, llama.cpp, and more. Demo: https://gpt.h2o.ai/ https://gpt-docs.h2o.ai/
Hi there, I installed everything with the CPU method on my Linux 20 server. After python generate.py --base_model =gptj --score_model=None I received: File "generate.py", line 50 load_8bit: bool = False,...
Test pipeline at: http://jenkins:8080/job/h2ogpt-test-draft
HI, I just entered "William Blabla is a developer" in the "Paste Text" field and selected MyData. Then I asked who William Blabla is and the answer was: William Blabla...
Dear Support Team, I recently upgraded my pip libraries, including transformers, peft, accelerate, and bitsandbytes, to support 4-bit training as opposed to the original 8-bit training. After doing so, I...
So i use this url: https://query1.finance.yahoo.com/v7/finance/download/AAPL?period1=345427200&period2=1685145600&interval=1d&events=history&includeAdjustedClose=true as data source. It is basically all the prices of the shares of apple. with the model: [Model: **h2oai/h2ogpt-oasst1-512-12b]** If i put this url...
While had added critical protection against failure when inputs too long, still if prompt happens to be too long, tokenization will do truncation=True for a max_length in some way (left...
Curious why no 4bit models. Would really amplify the ability to use larger models.
Here is a discussion: with model [Model: h2oai/h2ogpt-oasst1-512-12b] what are the necessary condition for a country to have the internet ? The necessary conditions for a country to have the...
Can someone provide me a step by step set of instructions (post installation) to use the software to make the summary of a text that is in a .txt file...
… of \n to , still outout in UI is wrong and all compressed. But somehow copy-paste of that text pops back to having the missing new lines. https://github.com/gradio-app/gradio/issues/4344