I hope qwen3 will open-source a large 120b model, like the GPT-120b. GPT's Chinese version isn't very good; we still need qwen.
#16 opened 5 days ago
by
mimeng1990
็ปๅ ธ้ป่พๆจ็้ขๆ ๆณ็ปๅบๆญฃ็กฎ็ญๆก
1
#15 opened 29 days ago
by
pypry
FP8 ็ๆฌๆ ๆณไฝฟ็จvllm v0.10.2 ๆฅๅ ่ฝฝ
#14 opened about 2 months ago
by
iwaitu
Good for Roleplay - Sucks at tool calling
#13 opened about 2 months ago
by
abiteddie
4 bit with bitsandbytes not working
#12 opened about 2 months ago
by
TheBigBlockPC
Fix broken qwen3-next blog link
#11 opened about 2 months ago
by
Smorty100
Something of with knowledge cutoff date?
#10 opened about 2 months ago
by
Nirav-Madhani
Failed to run VLLM batch inference
โ
1
1
#9 opened about 2 months ago
by
Junxiao-Zhao
๐โโ Best Practices for Evaluating the Qwen3-Next Model
๐
๐
7
1
#8 opened about 2 months ago
by
Yunxz
vllm: error: unrecognized arguments: --enable-reasoning
1
#7 opened about 2 months ago
by
evilll
awq
๐
8
#5 opened about 2 months ago
by
groxaxo
The Thinking model has a higher hallucination rate than the Instruct model and tends to overlook details in long contexts.
1
#4 opened about 2 months ago
by
xiaoxiao218
Local Installation Video and Testing On CPU - Step by Step
#3 opened about 2 months ago
by
fahdmirzac
Support of 1M context doubt
๐
1
1
#2 opened about 2 months ago
by
clyang33
GGUF when?
๐
๐
14
1
#1 opened about 2 months ago
by
ouchiewouchie