Dmytro Dzhulgakov

dzhulgakov's activity

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

New activity in deepseek-ai/DeepSeek-V3 2 months ago

Bug in fp8_cast_bf16.py

1

#4 opened 2 months ago by

dzhulgakov

New activity in meta-llama/Llama-3.2-11B-Vision-Instruct 5 months ago

Tokenizer needs to be fixed for BOS handling

#18 opened 5 months ago by

dzhulgakov

New activity in meta-llama/Llama-3.2-1B-Instruct 5 months ago

Tokenizer BOS behavior is inconsistent with Llama 3.1

1

#5 opened 5 months ago by

dzhulgakov

New activity in deepseek-ai/DeepSeek-Coder-V2-Instruct 8 months ago

How important is the grouped_topk?

#6 opened 8 months ago by

dzhulgakov

New activity in google/gemma-2-9b 8 months ago

Can't repro MMLU: sliding window attention implementation seems broken

3

#11 opened 8 months ago by

dzhulgakov

New activity in meta-llama/Meta-Llama-3-70B-Instruct 10 months ago

clean_up_tokenization_spaces=True causes formatting issues, why is it set?

2

#44 opened 10 months ago by

dzhulgakov

New activity in google/gemma-7b-it about 1 year ago

Running sample code gives ma a shape error

1

#22 opened about 1 year ago by

dzhulgakov

New activity in DiscoResearch/mixtral-7b-8expert about 1 year ago

Update modeling_moe_mistral.py

2

#1 opened about 1 year ago by

bjoernp

commented a paper over 1 year ago

Mistral 7B

Paper • 2310.06825 • Published Oct 10, 2023 • 46 •

8