render_message_to_json: Neither string content nor typed content is supported by the template. This is unexpected and may lead to issues.
2
#7 opened 3 days ago
by
whoisjeremylam
GLM 5.1 vs GLM 5 - burns A LOT output tokens on thinking
π 1
33
#6 opened 8 days ago
by
curiouspp8
Comparing against Unsloth UD_Q4_K_XL
π 1
3
#5 opened 8 days ago
by
TimothyRoo
Testing smol-IQ4_K
8
#4 opened 9 days ago
by
shewin
Testing IQ3_KS
π 1
10
#3 opened 9 days ago
by
shewin
Typical GLM 5.1 overhead on top of weights memory
π 1
4
#2 opened 10 days ago
by
curiouspp8
Fantastic as usual
β€οΈ 5
5
#1 opened 11 days ago
by
ndroidph