Rayen's picture

6 1 4

Rayen

Lissanro

https://Dragon.Studio

Lissanro

AI & ML interests

None yet

Recent Activity

new activity 2 days ago

Qwen/QwQ-32B:Is this model native 128K context length, or YaRN extended?

new activity 3 days ago

Qwen/QwQ-32B:Doesn't Generate `<think>` tags

liked a model about 1 month ago

turboderp/Qwen2.5-VL-7B-Instruct-exl2

View all activity

Organizations

None yet

Lissanro's activity

New activity in Qwen/QwQ-32B 2 days ago

Is this model native 128K context length, or YaRN extended?

#28 opened 4 days ago by

New activity in Qwen/QwQ-32B 3 days ago

Doesn't Generate `<think>` tags

#25 opened 4 days ago by

New activity in bigstorm/Llama-3.1-Nemotron-70B-Instruct-HF-7.0bpw-8hb-exl2 5 months ago

6bpw quant request

#2 opened 5 months ago by

New activity in turboderp/Llama-3-70B-Instruct-exl2 11 months ago

tokenizer_config.json and config.json specify wrong EOS token, causing the model to not function correctly in backends which do not read EOS tokens from generation_config.json

#1 opened 11 months ago by

New activity in LoneStriker/Meta-Llama-3-8B-Instruct-8.0bpw-h8-exl2 11 months ago

Works great on oobabooga, but always ends with assistant

#3 opened 11 months ago by

New activity in alpindale/WizardLM-2-8x22B 11 months ago

This could likely be dewokefied and possible even improved using mergekit's new 'Model Stock' method!

#5 opened 11 months ago by