Nico Hezel
Neiko2002
AI & ML interests
None yet
Recent Activity
liked a model about 15 hours ago
Tivaphraen/Geryon-9B-v1 new activity 1 day ago
deepreinforce-ai/Ornith-1.0-35B-FP8:FP8 seems to be broken liked a model 2 days ago
kasimat/Qwen3.6-27B-AEON-Ultimate-Uncensored-FP8-MTPOrganizations
FP8 seems to be broken
2
#1 opened 1 day ago
by
Neiko2002
Original ninja.template provides better results
3
#1 opened about 1 month ago
by
Neiko2002
comparison
17
#2 opened 2 months ago
by
kalle07
Broken config.json for vllm v0.21.0
#3 opened about 1 month ago
by
Neiko2002
Improved quality by changing the chat_template.jinja
4
#1 opened about 1 month ago
by
Neiko2002
tool calling?
1
#1 opened about 2 months ago
by
Neiko2002
tool calling?
1
#4 opened about 2 months ago
by
Neiko2002
tool calling?
1
#2 opened about 2 months ago
by
Neiko2002
Worse tool-calling accuracy due to chat_template.jinja
1
#2 opened about 1 month ago
by
Neiko2002
Crashes with newest vllm version (v0.20.1)
15
#1 opened about 2 months ago
by
Neiko2002
Does not work on 3090 GPUs
3
#2 opened about 2 months ago
by
Neiko2002
Amazing model
🔥 1
1
#3 opened about 2 months ago
by
Neiko2002
New activity in cyburn/Qwopus3.6-35B-A3B-v1-PrismaSCOUT-Blackwell-NVFP4-BF16-vllm-4.75bits about 2 months ago
Works on 3090
#1 opened about 2 months ago
by
Neiko2002
tool calls?
6
#4 opened 2 months ago
by
CryptoAIM
Removing speculative-config with care
#2 opened about 2 months ago
by
Neiko2002
RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (CUDABFloat16Type) should be the same
2
#4 opened over 1 year ago
by
Neiko2002
Flash Attention 2
2
#1 opened almost 2 years ago
by
Modularity