AryanEmbered

• Openai New Memory feature is just Vector Search?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 41 comments

• KBLaM by microsoft, This looks interesting

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 56 comments

• Is slower inference and non-realtime cheaper?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 5 comments

• No benchmarks or details on the performance of 0.6B qwen?🧐

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 12 comments

• Dolphin translator incoming (eventually)

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 6 comments

• A host of rumours

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 19 comments

• This Video model is like 5-8B params only? wtf

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 14 comments

• Llama 4 is not omnimodal

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 30 comments

• Google released Gemma 3 QAT, is this going to be better than Bartowski's stuff

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 32 comments

• Multi threaded LLM?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 8 comments

• Is a multimodal focused release from openai the best for us?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 20 comments

• Do you think this will catch on? Amazon's nova models are not very good.

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 12 comments

• Why don't we have LLMs that truly learn?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 27 comments

• How Much VRAM Do you need to run a 32B with 32k context?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 42 comments

• What is the current open SOTA for Text2SQL?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 2 comments

• What application to use Florence2 or SAM2 locally?

  Posted by AryanEmbered@reddit | LocalLLaMA | View on Reddit | 10 comments