Jonathan Schemoul

jon1012's activity

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

New activity in LibertAIDAI/GLM-5.3-Flash-NVFP4 24 days ago

Repetition

👍 1

5

#7 opened 2 months ago by

aeminkocal

New activity in LibertAIDAI/Qwen3.6-27B-NVFP4-GGUF 24 days ago

vision works fine with Qwen3.6-27B.mmproj-f32.gguf

1

#3 opened 5 months ago by

DGanno

New activity in LibertAIDAI/Gemma-4-12B-IT-NVFP4-GGUF 24 days ago

Will multimodal support exist in the future?

1

#1 opened 5 months ago by

Magnarks

New activity in LibertAIDAI/Nex-N2-mini-GGUF 24 days ago

This model is so good

1

#1 opened 5 months ago by

zmail-tech

New activity in LibertAIDAI/Qwen3.8-27B-NVFP4-GGUF 24 days ago

Thank you

❤️ 1

1

#1 opened 3 months ago by

josev814

New activity in LibertAIDAI/GLM-5.3-Flash-NVFP4 24 days ago

Datapoint: fp4 vs fp8 GLM-5.3-Flash endpoint parity on GSM8K + HumanEval

2

#8 opened 2 months ago by

joelafrite

Error with CUDA Graph Memory Calculations on T<4 set ups

1

#10 opened 2 months ago by

corgito

New activity in LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw 24 days ago

chat_template.jinja not fit Zig TensorFold

1

#5 opened 24 days ago by

JoeLeet

Reasoning effort dominates token cost in agentic use — worth a note on the card

3

#4 opened 25 days ago by

dpolist

Thank you for the quick upload

➕ 1

3

#1 opened 26 days ago by

joelafrite

Add chat_template.jinja (ported from DeepSeek's encoding.py via jayleaton's MIT pack)

1

#3 opened 25 days ago by

dpolist

New activity in LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw 25 days ago

Same-prefix reference forward request: exact-copy divergence on TP2 DGX Spark

1

#2 opened 25 days ago by

GeneDENG

New activity in LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-3.5bpw 25 days ago

Can't start.

1

#1 opened 25 days ago by

brewskybeer

New activity in nvidia/GLM-5.3-Flash-NVFP4 2 months ago

ignore / exclude_modules omits the MTP layer (45): checkpoint fails to load with speculative decoding

❤️ 4

1

#6 opened 2 months ago by

jon1012

New activity in LibertAIDAI/DeepSeek-V4.1-Flash-NVFP4 2 months ago

Thank you so much for your excellent job!!!!

1

#1 opened 2 months ago by

nakululusatuva

New activity in LibertAIDAI/GLM-5.3-Flash-NVFP4 2 months ago

B200 VLLM:glm53-flash

5

#2 opened 3 months ago by

corgito

The model's single-inference speed measured on two DGX systems is only 16 tokens/s, which does not meet productivity requirements.

1

#9 opened 2 months ago by

SAJDOQPDPS

GLM-5.3-Flash (NVFP4) on 4× DGX Spark — Multi-Node TP4 Benchmark (5.3–7.0 tok/s)

👍🔥 1

3

#5 opened 3 months ago by

tvearl

Deployment blocked: FlashInfer "Unsupported architecture" on DGX Spark (SM121)

4

#4 opened 3 months ago by

JayceHuang

GB10 tile values confirmed for DFLASH verify shape + a memory-config note for spec decoding

❤️ 1

1

#6 opened 3 months ago by

randomllama