Peter Chervenski

How do I use my custom dataset?

1

I don't want CIFAR, Taskonomy or anything, I have my own data, but I can't get it to work. I couldn't replace CIFAR10 and I don't know the data format...

ValueError: arrays must be the same length

2

I'm running the demo notebook unmodified on Ubuntu 20.04 machine with one GPU, everything is pip installed properly, but I keep getting this error below no matter what I do....

LlamaCpp model needs streaming support

1

Streaming is supported by llama-cpp-python and works in Jupyter notebooks outside langchain code, but I can't get it to work with langchain. I didn't see any code for streaming in...

It won't build on Windows

5

I keep getting this error, nothing builds: 6>C:\Users\spook\Dropbox\Desktop\ggml\examples\gpt-j\main.cpp(15,10): fatal error C1083: Cannot open include file: 'unistd.h': No such file or directory 5>C:\Users\spook\Dropbox\Desktop\ggml\examples\stablelm\main.cpp(15,10): fatal error C1083: Cannot open include file: 'unistd.h':...

Inference not working

3

I installed the diffusers package as specified on the HF page, and I use this code for simple inference: ```python import os import sys import torch from diffusers import StableCascadeDecoderPipeline,...

I couldn't install any OpenCV on QPython3. No package named opencv-aipy3 exists. Why is that?

I keep getting this list index out of range error for the example code: ========== IndexError Traceback (most recent call last) [](https://localhost:8080/#) in () 6 ae_best_valids, ae_best_tests = run_evolution_search(max_trained_models=length) 7...

Evaluation doesn't work on Windows

4

After getting a score of 0 every time, I looked at the samples.jsonl_results.jsonl file and the result for each is this: "failed: module 'signal' has no attribute 'setitimer'" This seems...

Code execution from Jupyter doesn't display matplotlib chart results like pyodide does

1

When this runs with Jupyter as the engine I get this:  But the chart displays normally with pyodide.

Execution timeout for Jupyter should be configurable

1

If a computation takes 10 seconds or longer with Jupyter (I didn't test pyodide but maybe the problem is present there as well) it fails:  This makes it hard...