Hugging Face's logo
Michael Royzen's picture
13 76

Michael Royzen PRO

michaelroyzen
21world's profile picture mcysqrd's profile picture Pent's profile picture
·

AI & ML interests

None yet

Organizations

veloce's profile picture Phind's profile picture

michaelroyzen's activity

New activity in Phind/Phind-CodeLlama-34B-v2 about 1 year ago

Wizardcoder controversy

19
#2 opened about 1 year ago by rombodawg

Issue/Bug replicating HumanEval result

9
#4 opened about 1 year ago by emrgnt-cmplxty

Phind-CodeLlama-13b model?

2
#3 opened about 1 year ago by rombodawg

I have some proof

1
#1 opened about 1 year ago by bhuwansaik
New activity in Phind/Phind-CodeLlama-34B-Python-v1 about 1 year ago

This might be the best coding model (Better than wizardcoder_python_34b)

5
#5 opened about 1 year ago by rombodawg

This is an absolute gem! Can't thank you enough.

12
#4 opened about 1 year ago by pgalko
New activity in Phind/Phind-CodeLlama-34B-v1 about 1 year ago

Python-only or multi-lingual?

1
#2 opened about 1 year ago by andersonbcdefg
New activity in Phind/Phind-CodeLlama-34B-Python-v1 about 1 year ago

True humaneval score or NewHope repeat?

5
#1 opened about 1 year ago by rombodawg
New activity in Phind/Phind-CodeLlama-34B-v1 about 1 year ago

instruct-answer format?

2
#1 opened about 1 year ago by theblackcat102
New activity in bigcode/the-stack-dedup about 1 year ago

Bug: datasets.builder.DatasetGenerationError: An error occurred while generating the dataset

3
#30 opened about 1 year ago by michaelroyzen
New activity in google/flan-t5-xxl over 1 year ago

Flan-T5 tokenizer supports neither Chinese nor many code-related tokens despite being advertised as such

1
#33 opened over 1 year ago by michaelroyzen
New activity in google/flan-t5-xxl almost 2 years ago

Only BF16 Work. FP16 and 8INT generate non-sense for me currently.

9
#9 opened almost 2 years ago by JulesGM

Activation function config inconsistent

6
#11 opened almost 2 years ago by michaelroyzen