coder543

20 ms·

451.
by coder543 3y ago
I literally paid for a month of Netflix in December, as I mentioned in the comment you replied to. I subscribed to Netflix for many years continuously before they massively inflated their prices. Yes, I would pay for it if their prices we
452.
by coder543 3y ago
I just want a 4K Originals Plan with 1 screen. Features: - Only Netflix Original content (in 4K) - 1 screen, since password sharing isn’t allowed anyways - No ads - No WWE - No smartphone games - No licensed conten
453.
by coder543 3y ago
I don’t see clear licensing on any of this, including the synthetic dataset.
454.
by coder543 3y ago
M1 was 68GB/s (and the article says this too). M2 was 100GB/s.
455.
by coder543 3y ago
gotcha, it’s cool that you’re able to get such good performance out of a WiFi 5 AP! And yep, OpenSpeedTest is nice!
456.
by coder543 3y ago
I agree your comment was relevant. The person I was responding to said they weren’t using WiFi 6 at all, which I interpreted to mean that all of their client devices were also still WiFi 5, but my interpretation could have been wrong. If yo
457.
by coder543 3y ago
Testing with a WiFi 5 client device (iPhone SE 1st gen), I'm seeing only 240Mbps down and 122Mbps up, which is far closer to what I would expect from WiFi 5. I don't know how a WiFi 6 client (like your phone) might work around the
458.
by coder543 3y ago
I checked Amazon, and the cheapest one I saw was ~$140, not <$100. I saw WiFi 6 (not 6E) options for <$100. Spending $200 now means that I don't have to spend $140 now and $140 again in two to three years.
459.
by coder543 3y ago
The UniFi U7 Pro that I bought was ~$200. For that money, it upgraded my network from WiFi 6 to WiFi 7, so now the 6E devices that I have are noticeably faster (since I didn't have 6E before), and my network is future-proofed for WiFi
460.
by coder543 3y ago
I have never seen a WiFi 5 device benchmark anywhere close to 866Mbps in real world scenarios. Most of the time, you're lucky to reach half of the "theoretical" numbers for WiFi due to the way that they're calculated.
461.
by coder543 3y ago
What do you like about it? Compared to GPT-3.5, Claude Instant seems to be the same or worse in quality according to both human and automated benchmarks, but also more expensive. It seems undifferentiated. And I would rather use Mixtral tha
462.
by coder543 3y ago
I'm not the person you're replying to, but that sentence comes from this research: https://arxiv.org/abs/2307.11760
463.
by coder543 3y ago
https://ollama.ai/library?sort=newest
464.
by coder543 3y ago
ChatGPT-3.5 isn’t even worth touching as an end-user application. Bard is better (due to having some integrations), but it’s still barely useful. ChatGPT-4 is on an another level entirely compared to either 3.5 or Bard. It is actually usefu
465.
by coder543 3y ago
ChatGPT-3.5 or ChatGPT-4? There is a big difference. For fun, I just asked ChatGPT-4 to generate a normalized database representation of vcard information: https://chat.openai.com/share/1c88813c-0a50-4ec6-ba92-4d6ff8...
466.
by coder543 3y ago
LibreOffice, MariaDB, Nextcloud, and OpenZFS might be a few prominent examples.
467.
by coder543 3y ago
With a proper grammar, you can require the "subject" field to be one of several valid entity names. In the prompt, you would tell the LLM what the valid entity names are, which room each entity is in, and a brief description of ea
468.
by coder543 3y ago
No. LLMs do not have memory like that (yet). Your 'scary' examples are very hypothetical and would require intentional design to achieve today; they would not happen by accident.
469.
by coder543 3y ago
The Celestia project was dormant for nearly a decade. It becomes hard to package unmaintained software. I haven’t kept up with who the new owner is or what they’re doing, so there could be reasons it hasn’t been picked up again by the distr
470.
by coder543 3y ago
Did you try Dolphin Phi-2? The Dolphin fine-tune seems better to me in this case.
471.
by coder543 3y ago
Just to report back on this in case anyone finds this comment in the future... standalone mode almost worked. On the UniFi U7 Pro, standalone mode was limited to 40MHz channel width on 5GHz and 160MHz channel width on 6GHz, which was a de
472.
by coder543 3y ago
Appears to be 18 months old, give or take: https://youtu.be/hS-NiaGeeVA
473.
by coder543 3y ago
"Mistral 7B" is "mistral-tiny", "Mixtral" is "mistral-small", and "mistral-medium" is something larger. https://mistral.ai/news/la-plateforme/
474.
by coder543 3y ago
Mistral Medium is available via their API, but not available for download, for example, so I find that confusing if you’re claiming their CEO claimed the plan is to be open for all models.
475.
by coder543 3y ago
How many tokens/s? Which quantization? If you could test Q4KM and Q3KM, it would be interesting to hear how the M2 Max does!
476.
by coder543 3y ago
That is a very interesting discussion. Weird to me that the quantization code wasn’t required to be in the same PR. Ika is also already talking about a slightly higher 2.31bpw quantization, apparently.
477.
by coder543 3y ago
ollama has been by far the easiest way for me, either on Linux directly (as I do now) or WSL2.
478.
by coder543 3y ago
I have not personally gotten to test things that way
479.
by coder543 3y ago
Mixtral is a larger model, so maybe that makes it more tolerant of that level of quantization? I’ve been impressed with 3-bit Mixtral, but I haven’t done a ton of side by sides against 4-bit because I haven’t felt the need.
480.
by coder543 3y ago
The fallback does seem to work! Although the 4-bit 7B models only run at 1 token every several seconds. I still wish Phi-2, Dolphin Phi-2, and TinyLlama-Chat-v1.0 were available, but I understand you have plans to make it easier to download

More ›