#qwen3 — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #qwen3, aggregated by home.social.
-
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses
https://quesma.com/blog/qwen38-27b-quantizations-benchmarked/
Comments: https://news.ycombinator.com/item?id=49611128
#HackerNews #Benchmarking #Qwen3.8 #27B #quantizations #4bit #1bit #AIperformance #Qwen38
-
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses
https://quesma.com/blog/qwen38-27b-quantizations-benchmarked/
Comments: https://news.ycombinator.com/item?id=49611128
#HackerNews #Benchmarking #Qwen3.8 #27B #quantizations #4bit #1bit #AIperformance #Qwen38
-
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses
https://quesma.com/blog/qwen38-27b-quantizations-benchmarked/
Comments: https://news.ycombinator.com/item?id=49611128
#HackerNews #Benchmarking #Qwen3.8 #27B #quantizations #4bit #1bit #AIperformance #Qwen38
-
Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses
https://quesma.com/blog/qwen38-27b-quantizations-benchmarked/
Comments: https://news.ycombinator.com/item?id=49611128
#HackerNews #Benchmarking #Qwen3.8 #27B #quantizations #4bit #1bit #AIperformance #Qwen38
-
RT @OrcaRouter: Jemand hat unabhängig 8 verschiedene Qwen3.8-27B-Modelle ohne Zensur oder Abliteration benchmarkt.
mehr auf Arint.info
#Abliteration #AIResearch #Benchmark #LLM #OrcaRouter #Qwen3 #arint_info
-
RT @OrcaRouter: Jemand hat unabhängig 8 verschiedene Qwen3.8-27B-Modelle ohne Zensur oder Abliteration benchmarkt.
mehr auf Arint.info
#Abliteration #AIResearch #Benchmark #LLM #OrcaRouter #Qwen3 #arint_info
-
RT @OrcaRouter: Jemand hat unabhängig 8 verschiedene Qwen3.8-27B-Modelle ohne Zensur oder Abliteration benchmarkt.
mehr auf Arint.info
#Abliteration #AIResearch #Benchmark #LLM #OrcaRouter #Qwen3 #arint_info
-
RT @OrcaRouter: Jemand hat unabhängig 8 verschiedene Qwen3.8-27B-Modelle ohne Zensur oder Abliteration benchmarkt.
mehr auf Arint.info
#Abliteration #AIResearch #Benchmark #LLM #OrcaRouter #Qwen3 #arint_info
-
RT @OrcaRouter: Jemand hat unabhängig 8 verschiedene Qwen3.8-27B-Modelle ohne Zensur oder Abliteration benchmarkt.
mehr auf Arint.info
#Abliteration #AIResearch #Benchmark #LLM #OrcaRouter #Qwen3 #arint_info
-
Lokale KI erstmals auf Cloud-Niveau?
-
Not a fan of Qwen 3.8 27B.
Overthinks too much => wasted tokens.
Has someone ran it without too much thinking? Does it perform well with that knob down, or letting it think all the way is required to be good?
#AI #ArtificialIntelligence #LLM #Qwen #Qwen38 #Qwen3 #Ollama #LlamaCCP #vLLM #LargeLanguageModels #LocalAI
-
Not a fan of Qwen 3.8 27B.
Overthinks too much => wasted tokens.
Has someone ran it without too much thinking? Does it perform well with that knob down, or letting it think all the way is required to be good?
#AI #ArtificialIntelligence #LLM #Qwen #Qwen38 #Qwen3 #Ollama #LlamaCCP #vLLM #LargeLanguageModels #LocalAI
-
Not a fan of Qwen 3.8 27B.
Overthinks too much => wasted tokens.
Has someone ran it without too much thinking? Does it perform well with that knob down, or letting it think all the way is required to be good?
#AI #ArtificialIntelligence #LLM #Qwen #Qwen38 #Qwen3 #Ollama #LlamaCCP #vLLM #LargeLanguageModels #LocalAI
-
Not a fan of Qwen 3.8 27B.
Overthinks too much => wasted tokens.
Has someone ran it without too much thinking? Does it perform well with that knob down, or letting it think all the way is required to be good?
#AI #ArtificialIntelligence #LLM #Qwen #Qwen38 #Qwen3 #Ollama #LlamaCCP #vLLM #LargeLanguageModels #LocalAI
-
Qwen3-next đã đạt tốc độ xử lý 103 token/giây trên CPU Intel Xeon v4 (22 lõi, 64GB RAM) mà KHÔNG CẦN GPU! Đây là một bước tiến đáng kể, cho phép chạy các mô hình ngôn ngữ lớn (LLM) hiệu quả hơn trên phần cứng CPU thông thường, đặc biệt với kiến trúc MoE và biên dịch Llama tối ưu.
#AI #LLM #CPU #Xeon #Qwen3 #NoGPU #Performance #TechNews #TríTuệNhânTạo #MôHìnhNgônNgữLớn #HiệuNăng #CôngNghệhttps://www.reddit.com/r/ollama/comments/1pb61u4/qwen3next_103_tokenssecond_in_cpu_xeon_v4_with_no/