COMMITS
/ exir/passes/_quant_patterns_and_replacements.py April 23, 2025
S
New embedding quant fusion
Scott Roy committed
January 28, 2025
S
Assert quant_min/quant_max in embedding4bit (#7410)
Scott Roy committed
October 15, 2024
S
Add 2b embedding op (#5800)
Scott Roy committed
July 31, 2024
E
Change deprecated impl_abstract to register_fake (#4392)
Erik Lundell committed
June 21, 2024
M
Add to_out_variant tests (#2880)
Manuel Candales committed
April 19, 2024
M
Fix embedding_4bit out variant (#3151)
Mengwei Liu committed
M
Fix quantized embedding export logic (#3095)
Mengwei Liu committed
April 18, 2024
M
Define embedding_4bit (#3121)
Manuel Candales committed
M
Delete llama_quantized lib (#3119)
Mengwei Liu committed
March 23, 2024
G
remove torchao dependency (#2599)
Guang Yang committed
March 15, 2024
M
use dequantize per channel group for embedding (#2374)
Manuel Candales committed
M
import more quantized decomposed ops into edge dialect (#2454)
Mengwei Liu committed
March 4, 2024
M
Add embedding_byte.dtype to enable output dtype be different than scales/zp dtype (#2236)
Manuel Candales committed
M
Update black linter in OSS lintrunner (#2229)
Mergen Nachin committed
M
Back out "Enable embedding_byte output dtype be different than scales/zp dtype" (#2210)
Manuel Candales committed
March 2, 2024
M
Enable embedding_byte output dtype be different than scales/zp dtype (#2091)
Manuel Candales committed
February 2, 2024
M
Add mixed mm op (#1791)
Manuel Candales committed
M
Define meta function for embedding_byte (#1793)
Michael Gschwind committed
January 31, 2024
M
Embedding with Half scales/output & null zero points (#1762)
Manuel Candales committed
August 11, 2023
S
Add pattern + replacement for Embedding with padding_idx
Salil Desai committed
July 31, 2023
F
Initial commit
facebook-github-bot committed