John

Lost messages in heavy duty websocket client ?

1

I am listening to hundreds of streams at a 100ms interval and it looks like ratchet starts to swallow and lose messages (or return them in wrong order) when the...

API is downloading outdated ca.pem, putting it in a wrong place and later not referencing it at all

2

The API downloads a file called ca.pem for SSL authentication by cURL. The idea is to solve SSL issues, in practice it will to the opposite. **Issue 1:** It puts...

txt2img GPU memory leak for batches -> crash after 5 hours with 22 GB consumed by PyTorch

1

**Describe the bug** I started a 2000 image batch production and it crashed quarter way through, there is a GPU memory loss * highres fix * euler a * 1024x768...

Not really useable, does not mask anything properly

3

I tested this quite a bit and while I appreciate the work and the idea behind it, this is not useful for real world usage. I've tested it on a...

Question: img2img alternative test - significantly less useful since the last updates (since sigma checkbox)

7

This is just a question to see if my personal experience matches with you guys. I've been using the img2img alternative test method quite successfully and frequently, it worked very...

Extremely low performance of Monterey (allocated 28 cores, 12 gig ram)

3

The mouse pointer is lagging behind 2-3 seconds on movements and clicks. It is not just the mouse, anything in the UI is slow, every menu is opening slow, just...

Just a suggestion: Think about what Automatic1111 did to Stable Diffusion

2

Think about what Automatic1111 did to Stable Diffusion, from a rather brute one-shot image generator significantly worse than the commercial counterparts it is now a distribution with thousands of features,...

This is just a sneaky advertisement for researchers to send their data to Meta.

3

Nice try. Like all other Meta "open" models and "open source" models it's the same game: You have to fill out one of their data collection portals, provide all details...

Something seems wrong with performance on Nvidia/cuda

6

### Describe the bug I have a 13th I9, 64gig ddr-5 ram and an idle RTX 3090 Fresh installed with anaconda Running llama 7B in 8 bit mode gives me...

Change ./main help output to better reflect context size's affect on generation length

2

### Discussed in https://github.com/ggerganov/llama.cpp/discussions/446 Originally posted by **cmp-nct** March 24, 2023 I've been testing alpaca 30B (-t 24 -n 2000 --temp 0.2 -b 32 --n_parts 1 --ignore-eos --instruct) I've consistently...