Hugging Face's logo
Jonna Matthiesen's picture
9 4 11

Jonna Matthiesen

JonnaMat
embedl
IceWander's profile picture branikita's profile picture robtacconelli's profile picture
Ā·

AI & ML interests

None yet

Recent Activity

reacted to AINovice2005's post with šŸ‘ about 2 hours ago
Pro tip: If you are finetuning any model with tensorboard logs enabled, be sure to upload them to HF Hub as event artifacts, they can be viewed instantly. šŸš€ I previously remembered this done in the notus model release: https://huggingface.co/argilla/notus-7b-v1/tensorboard Examples: https://huggingface.co/AINovice2005/ModernBERT-base-lora-cicflow-1m-r8/tensorboard https://huggingface.co/AINovice2005/ModernBERT-base-lora-cicflow-1m-r4/tensorboard https://huggingface.co/AINovice2005/ModernBERT-base-lora-cicflow-1m-r16 cc: @davidberenstein1957
replied to their post about 2 hours ago
šŸš€ FlashHead: Efficient Drop-In Replacement for the Classification Head in Language Model Inference šŸ”Ž Check out our latest FlashHead-enabled model: https://huggingface.co/embedl/Cosmos-Reason2-2B-W4A16-Edge2-FlashHead šŸ§© Seamless integration with vllm: ``` docker run --rm -it \ --network host \ --shm-size=8g \ --ulimit memlock=-1 \ --ulimit stack=67108864 \ --runtime=nvidia \ --name=vllm-serve \ -e HF_TOKEN=hf_*** \ -e HF_HOME=/root/.cache/huggingface \ embedl/vllm:latest-jetson-orin-flashhead \ vllm serve "embedl/Cosmos-Reason2-2B-W4A16-Edge2-FlashHead" \ --max-model-len 8192 \ --gpu-memory-utilization 0.75 \ --max-num-seqs 2 \ --trust-remote-code ```
published an article about 9 hours ago
FlashHead: Accelerating Language Model Inference ~ *Efficient drop-in replacement for the classification head*
View all activity

Organizations

Embedl's profile picture