gliptic

7 ms·

31.
by gliptic 2y ago
> while it's also fabricated in an older node Is this 7 nm node really older than raspberry pi 5's 16 nm node?
32.
by gliptic 2y ago
Where are you seeing Hyperbolic offering DeepSeek V3 API? I'm only seeing DeepSeek V2.5.
33.
by gliptic 2y ago
Apparently you don't believe in reading titles either.
34.
by gliptic 2y ago
Huffman still has a performance edge for static distributions. ANS bridges some of the performance gap between arithmetic coding and huffman.
35.
by gliptic 2y ago
Yes, the only exception is dictionaries used in preprocessing, but I think that's mostly a tradeoff to reduce the runtime.
36.
by gliptic 2y ago
This is especially true for instance when you have two or more tokens that are about equally likely, or one token that is virtually certain, which ranking would obscure.
37.
by gliptic 2y ago
This could be circumvented by _training_ the LLM on the fly on the previously observed file data. This is what Bellard's other NN compressor, nncp, does [1], which is currently #1 on Mahoney's benchmark [2]. Unfortunately this is
38.
by gliptic 2y ago
The model gives you a probability distribution over the tokens. You could use that directly with arithmetic coding, but there are ways to convert that to a distribution over e.g. the next byte instead which would improve efficiency further
39.
by gliptic 2y ago
I'm pretty sure it doesn't use ranking. That leaves a lot of performance on the table. Instead you would use the actual predicted token probabilities and arithmetic coding.
40.
by gliptic 2y ago
"You can tell the RL is done properly when the models cease to speak English in their chain of thought" -- Karpathy
41.
by gliptic 2y ago
Not D), because it says "The liar has at least one green hat." which isn't implied.
42.
by gliptic 2y ago
Maybe, but from what I've seen you run out of what theory can say about your non-crypto hash function pretty quickly.
43.
by gliptic 2y ago
Hash tests have always been targets. What else are you supposed to do for non-cryptographic hashes?
44.
by gliptic 2y ago
The weirdest thing about TeVeS IMO is that it adds additional fields that warp spacetime, so how is it not a dark matter theory?
45.
by gliptic 2y ago
Both an LLM and Stockfish would fail that test.
46.
by gliptic 2y ago
This is paranormal reincarnation mumbo-jumbo, not inheritance.
47.
by gliptic 2y ago
> I _think_ I understand the basic premise behind stable diffusion, i.e., reverse the denoising process to generate realistic images but, as far as I know, this is always done at the pixel level. It's typically not done at the pixel
48.
by gliptic 2y ago
The training cost is included because these NN compressors are running the training while they are compressing. They are general compressors, although the specific variants submitted here may be tuned a little to Wikipedia (e.g. the pre-pro
49.
by gliptic 2y ago
Zig doesn't (and never will) have operator overloading, so using any user-defined numeric types is awkward. It also already has vector types.
50.
by gliptic 2y ago
No, it wasn't. Crowdstrike denied it had to do with zeros in the files.
51.
by gliptic 2y ago
Azure having problems is not a black swan event.
52.
by gliptic 2y ago
> Choose any path you want for setting up first-party mode. Examples of paths you might want to use include: /metrics, /securemetric, /analytics, or preferably a random alphanumeric string that you don't use on your w
53.
by gliptic 2y ago
I mean without storing the freq table, which the last example you linked doesn't. Having to store a frequency table limits the encoder to do no better than the 0-th order entropy. Arithmetic coding can do arbitrarily complex prediction
54.
by gliptic 2y ago
Sorry, I (and others) have explained several times what an equivalent arithmetic coding scheme would have to do to make the same assumptions (and have the same limitations) as your scheme. It would have to assume there's _exactly_ N on
55.
by gliptic 2y ago
That isn't updating the counts after every symbol. Why don't you write one that does so you can compare them fairly?
56.
by gliptic 2y ago
> Whenever you're compressing a file on your computer the frequencies are known beforehand. Not at all. The best compressors learn to predict the frequencies as they go, they do not compute them up front and store them in the output
57.
by gliptic 2y ago
It's easy to see this octogonal wheel [rolling downhill] is faster than this round wheel [on a flat plain]!
58.
by gliptic 2y ago
Not if the counts are updated correctly.
59.
by gliptic 2y ago
> Huffman, ANS, arithmetic coding, etc do not include the frequency counts as part of the size afaik, though of course that needs to be stored to decode, if not please link me. They may or they may not. There are schemes doing both. Eith
60.
by gliptic 2y ago
With arithmetic coding this requires 1 bit to store: the value of the first bit. The other is known with 100% probability. This is very simple to implement, especially with bit-wise arithmetic coding. The problem is that the Shannon limit d

More ›