Neural nets often fail with (repetitive) gibberish output when the input is too different from the training data. This model appears to take in the entire text input at once or look ahead at the next input letters, so the unusual "bla bla" at the end can mess up outputs near the beginning.
Geoffrey Hinton has recently been talking about how analog and "imperfect" computing with specialized hardware/circuitry may yield much cheaper neural nets, that could easily be as large as human brains, but would only cost a few dollars and would be extremely cheap to run. Not a new idea, but it is a fairly promising outlook, I think.
But they have highly likely internal prototypes with higher bandwidth and latency. Also, with distilled latent diffusion one can probably generate text(-images) much faster anyhow as it could produce long chunks of text at once, rather than needing recurrently feed back the new token to the inputs.
Yes, but the losses in Figure 3 increase because the larger models see fewer data to keep the FLOP budget constant, not because of overfitting. Large models do not overfit very much, so the loss of a larger model will still be better compared to a smaller model when you keep dataset size constant.
Though isn't it highly likely that core devs working at the big tech giants have access to 10x-100x faster compute, e.g. some secret TPU successor at Google?
The loss decreases with greater model size at the same compute budget (i.e. stopping sooner regarding training data). Also some rehearsal/multi-epoch training improves the forgetting rate (thereby improving performance substantially), which hasn't been taken into account by Chinchilla et al. because they train <1 epoch.
Also, even with "Chinchilla laws", you still gain performance in a larger model, you just need a lot more data (if just as noisy) to reach the same level of convergence, but a larger model will have already partially converged to a superior model with the same amount data.
Yeah, continuous online learning by fine-tuning seems like an obvious way of making these models recall information from outside the perceptible context. One could also prompt the model to (recursively) summarize code and prepend this summary to each prompt, and/or enable the model to interactively query function definitions or code summaries before outputting a final answer (trained by RLHF). But any such tricks might also quickly be outcompeted by an even more general model, e.g. one that directly controls the GUI and can communicate with coworkers...
The model with the most similar name in this list is code-cushman-001 which is described as "Codex model that is a stronger, multilingual version of the Codex (12B) model in the paper".
The next stronger Codex model is called code-davinci-001 which appears to be a fine-tuned version of the GPT-3 Davinci model which is known to have 175B parameters. The model naming is alphabetical in the order of the model size:
Amazing if this is only a 12B model. If this already increases coding productivity by up to 50% (depending on kind of work), imagine what a 1T model will be capable of! I do wonder if some programmers at FAANG are already having access to a way more powerful coding assistants, and whether they code much at all at this point, or only make high level code specifications and then fix up the automatically generated code.
Google will not disappear. They already have much larger neural nets almost ready for deployment, plus they will be able to afford even larger ones in future. And size is all that matters while the techniques are mostly trivial.
My Firefox is now stuck with a note at the top of the screen saying something like "voxelchain.app is controlling your mouse cursor. Please press ESC to take over."
It is measurable, but not harmful to a meaningful extent. There are lots of sources of low-dose radiation in the natural human/primate/.../mammalian environment.
Speaking of which, is anyone aware of example code using the LSTM? I've been trying to get this to work, but there seems to be information missing e.g how to setup the input/output descriptors and how to manage input data: https://developer.apple.com/documentation/accelerate/bnns/us...