Eldar Kurtic's picture

14 4 2

Eldar Kurtic

ekurtic

·

AI & ML interests

Compression of deep neural networks

Recent Activity

updated a model about 1 month ago

neuralmagic/Sparse-Llama-3.1-8B-2of4

upvoted a paper about 2 months ago

"Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization

commented a paper about 2 months ago

"Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization

View all activity

Organizations

ekurtic's activity

commented a paper about 2 months ago

"Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization

Paper • 2411.02355 • Published Nov 4 • 46 •

New activity in neuralmagic/Llama-3.1-Nemotron-70B-Instruct-HF-FP8-dynamic about 2 months ago

Context length

#2 opened about 2 months ago by

New activity in open-llm-leaderboard/open_llm_leaderboard 3 months ago

Normalization for MMLU-Pro doesn't make sense

#947 opened 3 months ago by

New activity in TIGER-Lab/MMLU-Pro 3 months ago

Paper claims there are 10-choices but the test split has varying number from choices (anywhere from 3 to 10)

#22 opened 3 months ago by

New activity in open-llm-leaderboard/open_llm_leaderboard 3 months ago

Locally reproducible Leaderboard (v2) scores

#946 opened 3 months ago by

New activity in neuralmagic/Meta-Llama-3-8B-Instruct-FP8 4 months ago

Potential problematic behavior of truncation and/or padding?

#5 opened 4 months ago by

New activity in meta-llama/Meta-Llama-3-8B 7 months ago

Add `add_bos_token` and `add_eos_token`

#141 opened 7 months ago by

Why are "add_bos_token" and "add_eos_token" missing in tokenizer_config.json ?

#140 opened 7 months ago by

New activity in HuggingFaceTB/cosmopedia 9 months ago

Concat all subsets?

#10 opened 9 months ago by

New activity in glaiveai/glaive-coder-7b about 1 year ago

Replication of evals on HumanEval and MBPP

#1 opened about 1 year ago by

New activity in mosaicml/mpt-7b-chat about 1 year ago

Fix title in citation info

#35 opened about 1 year ago by

New activity in reciprocate/llama2-7b-gsm8k about 1 year ago

Finetuning recipe

#1 opened about 1 year ago by

New activity in tiiuae/falcon-7b over 1 year ago

Why is alibi: false in the config.json?

#69 opened over 1 year ago by