crunchg @crunchgee

Translation: Our model was generating cartoonish pictures that were useless anyway and not even close to what the prompt was. People were turning to grok and other one-click installation of offline solutions like "stability matrix". Our hypocrisy regarding "ethics" and "safety" -that was rendering our product into a total waste of time- had to walk away. People using smart prompting were jailbreaking our "safety" guardrails every time anyway... We regret wasting your time and causing all that frustration to people paying money to use our platform and now that the competition got tougher we realise that we need to stop pretending to be the holy inquisition.

7 11 105 23K 9

View Details

Akshay 🚀 @akshay_pachaar

a year ago

Let's fine-tune DeepSeek-R1 (distilled Llama) 100% locally:

21 202 2K 407K 4K

View Details

crunchg @crunchgee

a year ago

Everyone does this

Zuhayeer Musa @zuhayeer

a year ago

deepseek engineers are different 💀 to use deepseek's api, you “npm install openai” which is both genius and hilarious, saves so much work by mirroring their api and just swapping out the base url 😆 had to check out the docs to see this for myself, what a hack

317 524 8K 753K 2K

0 0 0 37 0

View Details

terminally onλine εngineer @tekbog

a year ago

i cant believe ChatGPT lost its job to AI

839 24K 214K 6.0M 6K

View Details

crunchg @crunchgee

a year ago

Cope

Mark Chen @markchen90

a year ago

Congrats to DeepSeek on producing an o1-level reasoning model! Their research paper demonstrates that they’ve independently found some of the core ideas that we did on our way to o1.

957 1K 22K 8.3M 5K

0 0 0 71 0

View Details

Jina AI @JinaAI_

a year ago

Dear Readers, you'll ❤️ this: Introducing ReaderLM-v2, a 1.5B small language model for HTML-to-Markdown conversion and HTML-to-JSON extraction with exceptional quality. Thanks to the new training paradigm and higher-quality training data, ReaderLM-v2 is a significant leap forward from its predecessor, particularly in handling long-context and markdown syntax. While the first generation approached HTML-to-markdown conversion as a "selective-copy" task, v2 treats it as a true translation process. This enables the model to masterfully leverage markdown syntax, excelling at generating complex elements like code fences, nested lists, tables and LaTex equations. You can use ReaderLM-v2 today via Reader API, HuggingFace, AWS Sagemaker, etc.

22 74 572 59K 512

View Details

crunchg @crunchgee

2 years ago

Don't do this ✍️✍️

Travis Arnold @souporserious

2 years ago

I was today years old when I realized you can just return a promise from a constructor to make it async. Constructors cannot use the async modifier so it's a nice way to do some async work before instantiation.

42 41 896 88K 398

0 0 0 20 0

View Details

crunchg @crunchgee

2 years ago

@JGragnier @docteur_TJ @this_is_peebs @elonmusk Says the French guy 😂😂😂

1 0 1 78 0

View Details

Vaibhav (VB) Srivastav @reach_vb

2 years ago

Fuck it! You can now run *any* GGUF on the Hugging Face Hub directly with @ollama 🔥 This has been a constant ask from the community, starting today you can point to any of the 45,000 GGUF repos on the Hub* *Without any changes whatsoever! ⚡ All you need to do is: ollama run hf. co/{username}/{reponame}:latest For example, to run the Llama 3.2 1B, you can run: ollama run hf. co/bartowski/Llama-3.2-1B-Instruct-GGUF:latest If you want to run a specific quant, all you need to do is specify the Quant type: ollama run hf. co/bartowski/Llama-3.2-1B-Instruct-GGUF:Q8_0 That's it! We'll work closely with Ollama to continue developing this further! ⚡

59 234 2K 318K 1K

View Details

Nick Dobos @NickADobos

2 years ago

Ai will significantly halt evolution of the human brain

241 36 604 278K 59

View Details

Jina AI @JinaAI_

2 years ago

Announcing reader-lm-0.5b and reader-lm-1.5b, jina.ai/news/reader-lm… two Small Language Models (SLMs) inspired by Jina Reader, and specifically trained to generate clean markdown directly from noisy raw HTML. Both models are multilingual and support a context length of up to 256K tokens. Despite their compact size, these models achieve state-of-the-art performance on this HTML2Markdown task, outperforming larger LLM counterparts while being only 1/50th of their size.

26 188 1K 157K 1K

View Details

crunchg @crunchgee

2 years ago

I have a concept of a plan

0 0 0 10 0

View Details

crunchg @crunchgee

2 years ago

v1.codehike.org/blog/remotion

0 0 0 15 0

View Details

Unsloth AI @UnslothAI

2 years ago

We made a step-by-step tutorial on how to finetune Llama-3 with Google Colab & deploy it to @Ollama Tutorial: docs.unsloth.ai/tutorials/how-… Colab notebook: colab.research.google.com/drive/1WZDi7AP… Blog post & video coming soon. 🦥