froggeric

froggeric's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in froggeric/Qwen-Fixed-Chat-Templates about 10 hours ago

Qwen 3.6 template has unnecessary cache invalidations

14

#1 opened 16 days ago by

sdroege

vLLM: 'Replace the chat_template string in your tokenizer_config.json with the file contents.' ... where?

4

#5 opened 7 days ago by

pathosethoslogos

When using Claude Code, tool calls end up broken with this chat template in Qwen3.6-27B

7

#6 opened 6 days ago by

mancub

think off , but answer in reasoning

1

#7 opened 6 days ago by

xd2333

Improvement (50% token reduction via the tool)

7

#8 opened 5 days ago by

Elsephire

`preserve_thinking` flag ignored?

#9 opened 5 days ago by

astride-thee-squid

Template V10 causes hallucinations.

5

#10 opened 5 days ago by

lorenzodimauro97

v11/v12 performance considerations with Claude Code?

3

#11 opened 4 days ago by

mancub

test v12 + qwen 27B + llamacpp + pi coding agent

2

#12 opened 4 days ago by

tuanzxcv

Abruptly stopping with <tool call>

2

#13 opened 3 days ago by

crotron

v13 stops dead after the first response

👍 2

5

#14 opened 3 days ago by

mancub

v15 released with fixes for overthinking/indecision and loops on tool errors

❤️ 3

4

#15 opened 1 day ago by

froggeric

v13 to v15 are completely broken in VLLM for qwen3_coder parser

5

#16 opened 1 day ago by

gudsjwixvfasptjhtr

Python_Search_file breaks the session

1

#17 opened 1 day ago by

Jokeru1983

v16 - Reasoning despite setting --reasoning off in llama.cpp

3

#18 opened about 19 hours ago by

diroka

updated a model about 10 hours ago

froggeric/Qwen-Fixed-Chat-Templates

Updated about 10 hours ago • 194

New activity in froggeric/Qwen-Fixed-Chat-Templates 7 days ago

Gemma 4 template request

3

#4 opened 7 days ago by

OnesimusTheLesser

New activity in froggeric/Qwen3.6-27B-MTP-GGUF 7 days ago

Thank You!

👍 1

1

#1 opened 9 days ago by

svarognsa

MTP output issues on Vulkan (6900 XT)

3

#2 opened 8 days ago by

SenorCiki

--no-mmproj-offload to avoid VRAM usage

🤝 1

2

#3 opened 8 days ago by

gdevenyi