🏗️ Building on HF

Marc Sun

marcsun13

huggingface

·

AI & ML interests

LLM, Quantization, Training, Inference

Recent Activity

upvoted an article 12 days ago

Is it agentic enough? Benchmarking open models on your own tooling

liked a Space 13 days ago

nanotron/ultrascale-playbook

liked a model 19 days ago

google/gemma-4-E4B-it-qat-mobile-transformers

View all activity

Organizations

New activity in kernels-community/triton_kernels 9 months ago

Implement make_opt_flags function for XPU

#5 opened 10 months ago by

Kernel assertion errors on 5090 using generation with MXfp4 (gpt-oss) - (stable on 4090)

#1 opened 11 months ago by

New activity in kernels-community/triton_kernels 10 months ago

Update torch-ext/triton_kernels/matmul_ogs_details/opt_flags.py

#3 opened 10 months ago by

Update build/torch-universal/triton_kernels/matmul_ogs_details/opt_flags.py

#2 opened 10 months ago by

New activity in openai/gpt-oss-20b 11 months ago

Don't think this works on 4090 with transformers, for me anyway

#53 opened 11 months ago by

New activity in openai/gpt-oss-120b 11 months ago

Request: 4-bit GPTQ or AWQ quantized version of openai/gpt-oss-20b

#32 opened 11 months ago by

running mxfp4 on H100 using tranformers with triton_kernel: make_default_matmul_mxfp4_w_layout not found

#64 opened 11 months ago by

gpt_oss error when running on kaggle

#72 opened 11 months ago by

New activity in openai/gpt-oss-20b 11 months ago

Open AI open source models not able to run ...

#34 opened 11 months ago by

RTX 5000 can't not use this model...

#41 opened 11 months ago by

MXFP4 only runs on h100 or b100 or later versions,

#61 opened 11 months ago by

information how to get it working on a 3090

#63 opened 11 months ago by

how to ignore the quantization

#70 opened 11 months ago by

New activity in openai/gpt-oss-120b 11 months ago

OOM on 3090

#52 opened 11 months ago by

New activity in hf-accelerate/accelerate_examples about 1 year ago

add multi_xpu configuration

#5 opened about 1 year ago by

New activity in hf-accelerate/model-memory-usage about 1 year ago

Error accessing Llama4 models: 'does not have any library metadata on the Hub'

#39 opened about 1 year ago by

GPTQ group size effects vRAM usage

#18 opened over 2 years ago by

Determining Minimum GPU Memory and Input Text Length Calculation in Model Training

#19 opened over 2 years ago by

microsoft/Florence-2-base raises error without any detail

#36 opened almost 2 years ago by

For Llama-2 model, I tried to get the memory usage, it says I do not have the access

#35 opened almost 2 years ago by