Osama_Saba

• Q3 is absolute garbage, but we always use q4, is it good?

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 24 comments

• People who don't enable flash attention - what's your problem?

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 46 comments

• What do we use for real time English speech recognition with low vram

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 16 comments

• Is vllm faster than ollama?

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

• Why no more progress in multimodals under 10b it's too slow I need something new or I sell my gpu not really joking but why

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 11 comments

• Qwen 14B is better than me...

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 362 comments

• Wait, can I abuse huggingface for storage for free??

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 13 comments

• Chached input locally?????

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 11 comments

• Wife running our local llama, a bit slow because it's too large (the llama not my wife)

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 74 comments

• Gpt 4o-mini vs models

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

• Lm studio makes the computer slow for no reason

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

• I don't want thinking models

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 15 comments

• I'm hungry for tool use

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

• I'm hungry for tool use

  Posted by Osama_Saba@reddit | LocalLLaMA | View on Reddit | 1 comments

• I want to know the reason for the go around but I can't understand what LOT 454 is saying

  Posted by Osama_Saba@reddit | aviation | View on Reddit | 19 comments