ALAE
gpt-2
Our great sponsors
ALAE | gpt-2 | |
---|---|---|
1 | 63 | |
3,492 | 20,917 | |
- | 1.8% | |
0.0 | 3.1 | |
about 3 years ago | 2 months ago | |
Python | Python | |
- | GNU General Public License v3.0 or later |
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.
ALAE
We haven't tracked posts mentioning ALAE yet.
Tracking mentions began in Dec 2020.
gpt-2
- Sam Altman is still trying to return as OpenAI CEO
- Build Personal ChatGPT Using Your Data
-
Don Knuth Plays with ChatGPT
Did you arrive at this certainty through reading something other than what OpenAI has published? The document [0] that describes the training data for GPT-2 makes this assertion hilarious to me.
[0]: https://github.com/openai/gpt-2/blob/master/model_card.md#da...
- По поводу опасности ИИ и предложений остановить разработки на 6 месяцев.
-
Any real competitor to GPT-3 which is open source and downloadable?
2.) OpenAI's GPT-2: OpenAI released GPT-2 in 2019 and later made the code and pre-trained models available to the public. GPT-2 is not as advanced as GPT-3 or GPT-4, but it's still powerful and can be used for various natural language processing tasks. You can find the GPT-2 code and pre-trained models on GitHub: https://github.com/openai/gpt-2
-
ChatGPT limitations
The Product is not the only implementation / available source for this technology / service / "thing". Given the amount of other LLMs that are available as trained models, OpenAI does not control access to this type of technology, in fact they OpenSourced their Previous model (https://github.com/openai/gpt-2). T5 (https://huggingface.co/docs/transformers/model_doc/t5) and Bloom (https://huggingface.co/docs/transformers/model_doc/bloom) are well known Generators with very comparable complexity to GTP-3. The fact is that OpenAI has the most reliable implementation, but far from the only implementation.
- The Concept Art Association updates their AI-restricting gofundme campaign, revealing their lack of AI understanding & nefarious plans! [detailed breakdown]
-
Anything like ChatGPT that you can run yourself?
You can try running GPT 2, I had some fun playing around with it, but it struggles with context. GPT 3 and now 3.5 blows it way out the water, but even if the model was available, you would need at least 350 GB of VRAM to run it. So unfortunately, for now it's out of reach.
-
OpenAI is fully Open-source? Will they ask for money?
The code for GPT-2 is here on github: https://github.com/openai/gpt-2
-
Deepfake Offensive Toolkit (real-time deepfakes for virtual cameras)
OpenAI attempted this with GPT-2’s incremental release strategy and was lambasted on moral and ethical grounds as well.
https://openai.com/blog/better-language-models/
So individuals working on a technology like this have no clear societal consensus to use as a guide when making the decision. Strong arguments can be made in both directions.
What are some alternatives?
dalle-mini - DALL·E Mini - Generate images from a text prompt
minGPT - A minimal PyTorch re-implementation of the OpenAI GPT (Generative Pretrained Transformer) training
Real-Time-Voice-Cloning - Clone a voice in 5 seconds to generate arbitrary speech in real-time
gpt-neo - An implementation of model parallel GPT-2 and GPT-3-style models using the mesh-tensorflow library.
sentencepiece - Unsupervised text tokenizer for Neural Network-based text generation.
jukebox - Code for the paper "Jukebox: A Generative Model for Music"
mesh-transformer-jax - Model parallel transformers in JAX and Haiku
gpt-neox - An implementation of model parallel autoregressive transformers on GPUs, based on the DeepSpeed library.
simsiam-cifar10 - Code to train the SimSiam model on cifar10 using PyTorch
dnn_from_scratch - A high level deep learning library for Convolutional Neural Networks,GANs and more, made from scratch(numpy/cupy implementation).
stylegan2-pytorch - Simplest working implementation of Stylegan2, state of the art generative adversarial network, in Pytorch. Enabling everyone to experience disentanglement