incredible things are happening in the diffusion land
@reach_vb
-
Basic Prompting Techniques Achieve Surprising Results in AI
By
–
This was a lovely read Max! Still surprised how far *just* basic prompting can take you.
-
QwQ 32B Draft Model Usage Discussion
By
–
I wonder if it’s more geared toward being used as a draft model for QwQ 32B
-
2025 AI Timeline Edition Released on Hugging Face
By
–
good call – this is how the 2025 edition will look like: https://
huggingface.co/spaces/reach-v
b/2025-ai-timeline
… does it look good to you? -
Updated LLM Models List: Hermes and Granite Added
By
–
Updated the list with more models – @NousResearch Hermes, IBM Granite, etc
-
DeepSeek v3: Cost-Effective GPT-4o Alternative for AI Apps
By
–
LocalLlama has spoken on DeepSeek v3 > If you used GPT-4o, you can safely switch; it’s the same thing at a much lower cost. Sometimes even better > v3 is the most ideal model for building AI apps. It is super cheap compared to other models, considering the performance
-
Phi 4 Model Availability: Azure to Hugging Face Migration
By
–
Phi 4 is only on Azure, they haven't uploaded it on HF yet, so I made a placeholder URL for when they upload lol. Will update it to use the Azure URL.
