5 ms·
Serious question: is there a warez scene for trained models yet? (I don't know how the model is accessed - are users of mainline GPT-3 given a .pb and a stack
by Mizza 6y ago
Serious question: is there a warez scene for trained models yet?
(I don't know how the model is accessed - are users of mainline GPT-3 given a .pb and a stack of NDAs, or do they have to access it through access-controlled API?)
Wherever data is desired by many but held by a few, a pirate crew inevitably emerges.
- exhilaration 6y agoIt's via API https://openai.com/blog/openai-api/ https://openai.com/blog/openai-api/
- jokowueu 6y agoI think this also might be an interest to you https://the-eye.eu/public/AI/pile_preliminary_components/ https://the-eye.eu/public/AI/pile_preliminary_components/
- MasterScrat 6y agoThose are datasets though, not models.
- jokowueu 6y agoI am aware I just thought you might find them interesting.
- notretarded 6y agoNot really
- kordlessagain 6y agoThe model is huge and is currently run in the cloud on many machines.
- mortehu 6y agoIt's only 175 billion parameters, so presumably it can fit on a single computer with 1024 GB RAM.
- Voloskaya 6y agoOn CPU the latency would be absolutly prohibitive to the point of being useless.
- typon 6y agoFor training yes, but not for inference.
- leogao 6y agoThe inference latency would also be prohibitive.
- Voloskaya 6y agoInference on GPU is already very slow on the full-scale non-distilled model (in the 1-2 sec range iirc), on CPU it would be an order of magnitude more.
- kordlessagain 6y agoFrom 2019: https://heartbeat.fritz.ai/deep-learning-has-a-size-problem-ea601304cd8 https://heartbeat.fritz.ai/deep-learning-has-a-size-problem-... > Earlier this year, researchers at NVIDIA announced MegatronLM, a massive transformer model with 8.3 billion parameters (24 times larger than BERT) > The parameters alone weigh in at just over 33 GB on disk. Training the final model took 512 V100 GPUs running continuously for 9.2 days. Running this model on a "regular" machine at some useful rate is probably not possible at this time.
- woko 6y agoSorry, but I don't see the link between the quote and your sentence.
- Voloskaya 6y agoCheckpoint is not shared with customers, you only get access to an API endpoint.
- vessenes 6y agoGPT-3 users are given an API link which routes to Azure, full blackbox.