Steve Li's picture

Steve Li

CHNtentes

·

CHNtentes

AI & ML interests

None yet

Recent Activity

new activity 2 days ago

unsloth/Qwen3-30B-A3B-GGUF:UD version for the Q5, Q6 and Q8 quant

liked a model 5 days ago

unsloth/Qwen3-30B-A3B-GGUF

new activity 8 days ago

ibm-granite/granite-4.0-tiny-preview:Speed comparison?

View all activity

Organizations

None yet

CHNtentes's activity

New activity in unsloth/Qwen3-30B-A3B-GGUF 2 days ago

UD version for the Q5, Q6 and Q8 quant

#11 opened 6 days ago by

New activity in ibm-granite/granite-4.0-tiny-preview 8 days ago

Speed comparison?

#3 opened 8 days ago by

New activity in Qwen/Qwen3-30B-A3B 8 days ago

Qwen3 is great, but could be better.

#18 opened 9 days ago by

New activity in unsloth/Qwen3-30B-A3B-GGUF 12 days ago

UPDATE FIXED! Is lower quants usable now?

#4 opened 12 days ago by

UPDATED FIXED!! Template problem?

#2 opened 12 days ago by

Noonecares52647842

New activity in unsloth/GLM-4-32B-0414-GGUF 13 days ago

Is UD-Q5_K_XL better than Q5_K_M one?

#3 opened 13 days ago by

New activity in unsloth/gemma-3-12b-it-qat-GGUF 13 days ago

Are these better than regular (non-qat) ones?

#1 opened 13 days ago by

New activity in bartowski/THUDM_GLM-4-32B-0414-GGUF 14 days ago

llama.cpp fixes have just been merged

#5 opened 17 days ago by

New activity in bartowski/THUDM_GLM-4-32B-0414-GGUF 19 days ago

Broken results

#2 opened 26 days ago by

New activity in moonshotai/Kimi-VL-A3B-Thinking 25 days ago

gguf?

#1 opened about 1 month ago by

AlgorithmicKing

New activity in city96/HiDream-I1-Full-gguf 25 days ago

Can we use existed llama 3.1 gguf

#1 opened 25 days ago by

New activity in meta-llama/Llama-4-Scout-17B-16E-Instruct about 1 month ago

Unethical comparisons with Deepseek replacing chinese languages by thai/vietnamese only

#32 opened about 1 month ago by

New activity in google/gemma-3-27b-it about 1 month ago

SigLIP or SigLIP2 encoder?

#48 opened about 1 month ago by

New activity in deepseek-ai/DeepSeek-V3-0324 about 1 month ago

Downloading weights without duplicates

#52 opened about 1 month ago by

听我说谢谢你因为有你温暖了四季

#37 opened about 2 months ago by

New activity in Qwen/Qwen2.5-Omni-7B about 1 month ago

Can this be quantized with bitsAndBytes?

#6 opened about 2 months ago by

New activity in Qwen/Qwen1.5-MoE-A2.7B-Chat-GPTQ-Int4 about 2 months ago

Int4为什么比没量化的float32和float16还慢

#3 opened 2 months ago by

New activity in THUDM/CogView4-6B about 2 months ago

Maybe rename it 15B

#5 opened 2 months ago by

New activity in mistralai/Mistral-Small-3.1-24B-Instruct-2503 about 2 months ago

Quantized models with vision included?

#27 opened about 2 months ago by

New activity in unsloth/QwQ-32B 2 months ago

I could be wrong, but I think the <think> tag needs to be removed from the last bit of the jinja template in the tokenizer_config.json

#1 opened 2 months ago by