294 19 344

John Leimgruber III PRO

ubergarm

https://www.paypal.com/donate/?hosted_button_id=HU59345BZVSUA

AI & ML interests

Open LLMs and Astrophotography image processing.

Recent Activity

new activity about 7 hours ago

ubergarm/GLM-4.7-GGUF:Stable run on 2x RTX 5090 and 2 Xeon E5 2696 V4 and DDR4 with ik_llama.cpp - 6.1 t/s on IQ4_K and 5.1 t/s on IQ5_K, opencode works with this

new activity 2 days ago

ubergarm/GLM-4.7-Flash-GGUF:question about mxfp4

new activity 2 days ago

zai-org/GLM-4.7-Flash:Why does the KV cache occupy so much GPU memory?

View all activity

Organizations

New activity in ubergarm/GLM-4.7-GGUF about 7 hours ago

Stable run on 2x RTX 5090 and 2 Xeon E5 2696 V4 and DDR4 with ik_llama.cpp - 6.1 t/s on IQ4_K and 5.1 t/s on IQ5_K, opencode works with this

👍 1

#5 opened 28 days ago by

martossien

New activity in ubergarm/GLM-4.7-Flash-GGUF 2 days ago

question about mxfp4

#3 opened 2 days ago by

koifish12

New activity in zai-org/GLM-4.7-Flash 2 days ago

Why does the KV cache occupy so much GPU memory?

#21 opened 3 days ago by

yyg201708

New activity in ubergarm/GLM-4.7-Flash-GGUF 2 days ago

Re-cooking imatrix and quants with updated ik/llama.cpp PR

🚀 1

#1 opened 3 days ago by

ubergarm

New activity in zai-org/GLM-4.7-Flash 3 days ago

Cannot run vLLM on DGX Spark: ImportError: libcudart.so.12

#18 opened 4 days ago by

yyg201708

Performance Discussion

👀 2

#1 opened 4 days ago by

IndenScale

Enormous KV-cache size?

👍 ➕ 4

#3 opened 4 days ago by

nephepritou

New activity in noctrex/GLM-4.7-Flash-MXFP4_MOE-GGUF 3 days ago

Feedback from running in LM Studio 0.39.3 with v1.103.2 of llama.cpp

#1 opened 3 days ago by

spanspek

New activity in ArtusDev/requests-exl 5 days ago

[QUANTING UPDATE]

❤️ 👍 3

#28 opened 8 days ago by

ArtusDev

New activity in ubergarm/Devstral-Small-2-24B-Instruct-2512-GGUF 5 days ago

Mistral 3 large wuant

👍 1

#1 opened 6 days ago by

facedwithahug

New activity in ubergarm/DeepSeek-V3.2-Speciale-GGUF 5 days ago

QuIP - 2 bit quantised as good as 16 bit

#5 opened 10 days ago by

infinityai

New activity in msievers/gemma-3-1b-it-qat-q4_0-gguf 9 days ago

Thanks for sharing your work!

❤️ 2

#1 opened 10 days ago by

ubergarm

New activity in ubergarm/DeepSeek-V3.2-Speciale-GGUF 9 days ago

Say Whattt?!

🔥 👍 4

#1 opened 15 days ago by

mtcl

New activity in ubergarm/Devstral-2-123B-Instruct-2512-GGUF 10 days ago

Decent PPL with 100% IQ4_KSS

🔥 1

#3 opened about 1 month ago by

sokann

New activity in kyutai/pocket-tts 10 days ago

Open access to the model

#1 opened 10 days ago by

jujutechnology

New activity in ubergarm/DeepSeek-V3.2-Speciale-GGUF 10 days ago

gguf_init_from_file_impl: tensor 'token_embd.weight' has invalid ggml type 139 (NONE)

#3 opened 13 days ago by

issa54

New activity in ubergarm/DeepSeek-V3.2-Speciale-GGUF 11 days ago

IQ4_K, etc.?

👀 1

#4 opened 11 days ago by

matdef99

New activity in sszymczyk/DeepSeek-V3.2-nolight-GGUF 12 days ago

Feedback

#1 opened 17 days ago by

bibproj

New activity in ubergarm/DeepSeek-V3.2-Speciale-GGUF 13 days ago

Testing IQ5_K

🤗 1

#2 opened 13 days ago by

shewin

New activity in deepseek-ai/DeepSeek-V3.2-Speciale 15 days ago

GGUF quants?

#17 opened 25 days ago by

fsaudm

John Leimgruber III PRO

AI & ML interests

Recent Activity

Organizations

ubergarm's activity

Stable run on 2x RTX 5090 and 2 Xeon E5 2696 V4 and DDR4 with ik_llama.cpp - 6.1 t/s on IQ4_K and 5.1 t/s on IQ5_K, opencode works with this

question about mxfp4

Why does the KV cache occupy so much GPU memory?

Re-cooking imatrix and quants with updated ik/llama.cpp PR

Cannot run vLLM on DGX Spark: ImportError: libcudart.so.12

Performance Discussion

Enormous KV-cache size?

Feedback from running in LM Studio 0.39.3 with v1.103.2 of llama.cpp

[QUANTING UPDATE]

Mistral 3 large wuant

QuIP - 2 bit quantised as good as 16 bit

Thanks for sharing your work!

Say Whattt?!

Decent PPL with 100% IQ4_KSS

Open access to the model

gguf_init_from_file_impl: tensor 'token_embd.weight' has invalid ggml type 139 (NONE)

IQ4_K, etc.?

Feedback

Testing IQ5_K

GGUF quants?