Oranguru

8 ms·

1.
by Oranguru 5d ago
You are too deeply mired in propaganda for any meaningful debate; you should read more neutral, international news sources to get a clearer picture. This was all about oil; not drugs, not terrorism, not communism. One less criminal is nothi
2.
by Oranguru 5d ago
AI has not yet reached the point where it can convincingly generate long presentations with the precision required for complex, technical academic topics. It will eventually get there, at which point an additional layer of proof of work wil
3.
by Oranguru 15d ago
Nevertheless, Venezuela remains in the same situation as before. Aside from Maduro, the corrupt and criminal government remains intact, the head of state handpicked by the US president, no public elections in sight. There has been no politi
4.
by Oranguru 1mo ago
> We're all more likely to get what we want. Ideally peace and prosperity. As sovereign citizens of their own country, Ukrainians have the right to decide under which flag they want to collectively pursue peace and prosperity. I bel
5.
by Oranguru 1mo ago
This is different now. US labs and companies are not releasing frontier-level models openly (specially those capable of assisting cyber intelligence work), but commercializing them instead. Thus, any ban would not be symmetrical to begin wi
6.
by Oranguru 3mo ago
They didn't say that; they simply said this is not the way of democracy. Legality has nothing to do with it.
7.
by Oranguru 3mo ago
But this isn't a story, literature, or a fancy piece of art; it's merely a technical blog post that discloses a security vulnerability. Here, the writing serves only as a vehicle to convey a message. Once you've received it,
8.
by Oranguru 7mo ago
Prompt injection is still a possibility, so while it improves the security posture, not by much.
9.
by Oranguru 10mo ago
What "other platforms" are you talking about? Just an Android app would suffice. It's not a huge deal for a company worth trillions, especially if the features are already there and they're just blocking non-Apple produc
10.
by Oranguru 1y ago
You can absolutely sync your vault without a paid subscription. Simply save it within your OneDrive or Google Drive folder. Alternatively, you could use Syncthing if you prefer a self-hosted solution.
11.
by Oranguru 2y ago
However, these rights should be guaranteed to a company operating in the USA and strictly adhering to US law. Of course, if the law is (arbitrarily) changed to make this illegal due to the Chinese government's stake, then it could be f
12.
by Oranguru 2y ago
Because both of these are based on Chromium (the open-source version of Chrome).
13.
by Oranguru 2y ago
This approach is valuable because it abstracts away certain complexities for the user, allowing them to focus on the code itself. I found it especially beneficial for users who are not willing to learn functional languages or parallelize co
14.
by Oranguru 2y ago
You can easily fix this using a grammar constraint with llama.cpp. Add this to the command: --grammar "root ::= [^一-鿿ぁ-ゟァ-ヿ가-힣]*" This will ban Chinese characters from the sampling process. Works for Yi and Qwen models.
15.
by Oranguru 3y ago
You can access GPUs within containers using CDI (Container Device Interface): https://docs.nvidia.com/datacenter/cloud-native/container-to... No additional tools (e.g., nvidia-ctk) are needed. Docker has recently
16.
by Oranguru 3y ago
Very interesting. Thanks for the references. Have you released the code or pre-trained models yet or do you plan to do so at some point?
17.
by Oranguru 3y ago
Thank you for your work. I have been having trouble achieving the desired level of formality in the generated text. When I ask for slightly formal content, the result tends to be too formal. However, when I ask the model to reduce the forma
18.
by Oranguru 3y ago
This comment does nothing more than repeat the information provided in the title and in a very verbose style. I don't understand the tendency for GPT generated comments that is lately taking over HN and Twitter threads. They are insipi
19.
by Oranguru 3y ago
It's about minimising friction in our tasks and reducing any unnecessary obstacles. Even seemingly minor actions that only take a few seconds can build up over time and generate a sense of frustration, especially when they become frequ
20.
by Oranguru 3y ago
Great job! In the future, would you add support for local LLMs, such as LLaMa?
21.
by Oranguru 3y ago
Yes, check out: https://huggingface.co/chat/ You can easily opt out of the data sharing.
22.
by Oranguru 3y ago
Useless for what? Are you comparing the base model with chat-tuned models? Chat-tuned derivatives of LLaMa 2 are already appearing. Given that the base LLaMa 2 model is more efficient than LLaMa 1, it is reasonable to expect that these more
23.
by Oranguru 3y ago
"Compared to GPT2 it’s on par" Any benchmarks or evidence tu support this claim? IF you try to find them, official benchmarks will tell you that this is not true. Even the smallest LLaMa model (7B) is far ahead of GPT2, like an or
24.
by Oranguru 3y ago
This is also a requirement for the research conducted under a project funded by the european comission.
25.
by Oranguru 3y ago
Take a look at: https://huggingface.co/blog/trl-peft
26.
by Oranguru 3y ago
Yes, there are. For example: https://laion.ai/blog/open-flamingo/
27.
by Oranguru 3y ago
I must remind you that large language models are not designed to perform arithmetic calculations nor they have been trained to do so. They are trained to recognize patterns in large amounts of text data and generate responses based on that
28.
by Oranguru 4y ago
Hallucinating? you mean overfitting, right?
29.
by Oranguru 4y ago
For STT, take a look at Wenet: https://github.com/wenet-e2e/wenet They provide support for running in a Raspberry Pi and it runs in real-time. I have tried the desktop version and the quality is good enough when the au
30.
by Oranguru 4y ago
This technical report from the European Commission was very useful in the research I conducted on this topic earlier this year. https://publications.jrc.ec.europa.eu/repository/handle/JRC1...

More ›