DevTools VS GLM-130B

Constantly being told to "Reload DevTools" · Issue #145 · MicrosoftEdge/DevTools (github.com)

If you've got specific feedback they have a GitHub repo - https://github.com/MicrosoftEdge/DevTools/issues/

I've published an explainer on GitHub that covers possible scenarios and will get more design ideas soon and we have an Issue on GitHub tracking the progress. I would love to hear more what you think would be a better experience than what we have right now.

Hey, Reddit friends! We've had a crazy couple of weeks here on the Edge team, which messed up a couple of releases. We should be back to our regular schedule now. If you missed it, we announced how to use Microsoft Edge Beta on your Steam Deck to use Xbox Cloud Gaming! This was super fun to partner with Xbox on, and if you have a Steam Deck we'd love to know about your experience. We had a few people run into issues with games freezing and needing to clear cache to continue but know that we are working on resolving that. I will update my announcement post with more details once I have those to share. We also launched a GitHub repo for feedback around the F12 DevTools, and posted up details on how the auto alternative text descriptions for images works!

Today we released a new GitHub repository for the Developer Tools in Microsoft Edge where you can file issues about problems, bugs and feattures you would like to see. You can also view the list of existing issues and add your comments.

The https://github.com/THUDM/GLM-130B model is trained on The Pile and can run on 4x3090 when quantized to INT4. I'm wondering if anyone knows if this model could (or has) been quantized using GPTQ, which gives some impressive performance gains over traditional quantization, and I'm also wondering if anyone has tried a 3-bit or 2-bit quantization of such a massive model (using GPTQ). Are there any inherent limitations in this? Is there anything about this model that prevents it from being run on text-generation-webui?

> GLM-130B, a model comparable with GPT-3, has 130 billion parameters in FP16 precision, a total of 260G of GPU memory is required to store model weights. The DGX-A100 server has 8 A100s and provides an amount of 320G of GPU memory (640G for 80G A100 version) so it suits GLM-130B well.

https://github.com/THUDM/GLM-130B/blob/main/docs/low-resourc...

GLM-130B[1] (a 130 billion parameter model vs GPT-3's 175 billion parameter model) is able to run optimally on consumer level high-end hardware, 4xRTX 3090 in particular. That's < $4k at current prices, and as hardware prices go one can only imagine what it'll be in a year or two. It also enables running with degraded performance on lesser systems.

It's a whole lot cheaper to run neural net style systems than to train them. "Somebody on Twitter"[2] got it setup, and broke down the costs, demonstrated some prompts, and what not. Cliff notes being a fraction of a penny per query, with each taking about 16s to generate. The output's pretty terrible, but it's unclear to me whether that's inherent or a result of priority. I expect OpenAI spent a lot of manpower on supervised training, whereas this system probably had minimal, especially in English (it's from a Chinese university).

[1] - https://github.com/THUDM/GLM-130B

[2] - https://twitter.com/alexjc/status/1617152800571416577

GLM-130B in 4 bit mode is better than GPT3 and can run on 4 RTX-3090s. Still expensive but it’s getting closer. https://github.com/THUDM/GLM-130B

https://github.com/THUDM/GLM-130B might be a useful place to look