GLM-130B
homepage
GLM-130B | homepage | |
---|---|---|
19 | 181 | |
7,616 | 16,141 | |
0.4% | 8.7% | |
4.8 | 9.9 | |
10 months ago | 7 days ago | |
Python | JavaScript | |
Apache License 2.0 | GNU General Public License v3.0 only |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
GLM-130B
-
GLM-130B
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?
- Has anyone tried GLM?
- Ask HN: Open source LLM for commercial use?
- Whichever way I look at it, I just don’t see this being the case. Why do you agree/disagree?
-
The New Bing and ChatGPT
> 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...
-
OpenAI Major Outage
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
- [D]Are there any known AI systems today that are significantly more advanced than chatGPT ?
-
Will there ever be a "Stable Diffusion chat AI" that we can run at home like one can do with Stable Diffusion? A "roll-your-own at home ChatGPT"?
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
- Open-Source competitor to OpenAI?
-
Ask HN: Can you crowdfund the compute for GPT?
https://github.com/THUDM/GLM-130B might be a useful place to look
homepage
- Highly customizable homepage with Docker and service API integrations
- Homepage JDownloader widget
-
Just started building a home server in my Raspberry Pi 3B+
It's Homepage. It's great for dashboarding, but has a few shortcomings in that you need to secure it behind a reverse proxy, otherwise you'll end up leaking credentials to the whole internet, unless you abstain from using its "connectors".
-
Just started homelabbing in an old Raspberry Pi 3B+
I use dietpi as os, the dash board is from homepage
- Bookmark manager with a focus on organization?
- Is there a dashboard to list the services I have running?
-
Dashboard for monitoring
I use Homepage. Has integrations with nearly every service I use and it's pretty easy to set up
-
Setting up a local domain
Step 2. Build a Dashboard. There are many options for personal dashboards, but I run Ben Phelps' Homepage in a Docker container. It is fast and simple to configure with YAML files. Here is a screenshot of my home dashboard. Homepage has more features than I use. Any ports needed for your services will be added to the URLs in the Homepage config file. Then, all you need to do is create a bookmark to Homepage in your partner's browser.
-
It's dashboard Wednesday! And I'm finally content with how mine looks;)
Good to see a dashboard post here that isnt just using Homepage :)
-
What kind of Alpine user are you?
The control panel is called Homepage. I like it more than Heimdall. To manage Docker I use Portainer.
What are some alternatives?
PaLM-rlhf-pytorch - Implementation of RLHF (Reinforcement Learning with Human Feedback) on top of the PaLM architecture. Basically ChatGPT but with PaLM
Heimdall - An Application dashboard and launcher
ggml - Tensor library for machine learning
homer-dashboard
petals - 🌸 Run LLMs at home, BitTorrent-style. Fine-tuning and inference up to 10x faster than offloading
Organizr - HTPC/Homelab Services Organizer - Written in PHP
Open-Assistant - OpenAssistant is a chat-based assistant that understands tasks, can interact with third-party systems, and retrieve information dynamically to do so.
homarr - Customizable browser's home page to interact with your homeserver's Docker containers (e.g. Sonarr/Radarr)
lm-human-preferences - Code for the paper Fine-Tuning Language Models from Human Preferences
Portainer - Making Docker and Kubernetes management easy.
hivemind - Decentralized deep learning in PyTorch. Built to train models on thousands of volunteers across the world.
Speedtest-Tracker - Continuously track your internet speed