dreamai87

• PI agent integrated with Cline-Kanban repo: All using PI and Qwen 3.6 35B MOE UD 4K_XL

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 10 comments

• AutoML pipeline one-shot experiment with Qwen 3.6 35B (llama.cpp) and PI coding agent

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 6 comments

• Qwen 3.6 35 UD 2 K_XL is pulling beyond its weight and quantization (No one is GPU Poor now)

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 68 comments

• Qwen3.5 35b is sure still one the best local model (pulling above its weight) - More Details

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 52 comments

• Comparison Qwen 3.6 35B MoE vs Qwen 3.5 35B MoE on Research Paper to WebApp

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 29 comments

• Qwen3.5 35b is sure one the best local model (pulling above its weight)

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 50 comments

• Just wanna say : Kudos to llama cpp our unsung heroes 🫡

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 4 comments

• Closed-Source AI Strikes Again: Cheap Moves Like This Prove We Need Open-Source Alternatives

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 42 comments

• No offense: Deepseek 8b 0528 Qwen3 Not Better Than Qwen3 8B

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 31 comments

• Flux.1 converted into GGUF - what interesting opportunity it offers in llm space?

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 66 comments

• Is it possible to convert fp8 safetensor to int4 or fp4

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 4 comments

• Test LLM on This Riddle

  Posted by dreamai87@reddit | LocalLLaMA | View on Reddit | 5 comments