llama.cpp
Introduce bfloat16 support
#6412
Merged

Introduce bfloat16 support #6412

ggerganov merged 8 commits into ggml-org:master from jart:bf16
jart
jart jart force pushed 1 year ago
jart jart force pushed 1 year ago
jart jart force pushed 1 year ago
github-actions
JohannesGaessler
sorasoras
jart
JohannesGaessler
jart
JohannesGaessler
1aienthusiast
jart
sorasoras
jart
Artefact2
JohannesGaessler
cpumaxx
sorasoras
cpumaxx
sorasoras
cpumaxx
jart jart force pushed to 07cebab5 1 year ago
jart
jart
cpumaxx
jart
jart
cpumaxx
jart
cpumaxx
cpumaxx
cpumaxx
jart jart force pushed from 07cebab5 1 year ago
jart jart force pushed 1 year ago
ggerganov
ggerganov approved these changes on 2024-04-09
jart jart force pushed 1 year ago
jart
jart
jart
ryao
JohannesGaessler
ggerganov
ggerganov approved these changes on 2024-04-25
ddh0
jart jart force pushed 1 year ago
jart
github-actions
jart jart force pushed to 68614cec 1 year ago
jart jart force pushed from ed0f47b3 to 82aebcf0 1 year ago
unicomp21
Srihari-mcw
Srihari-mcw
jart
jart Introduce bfloat16 support
55e962a2
jart Remove GGML code that's not needed
823d45ad
jart Minimize the GGML API surface area for BF16
180bfcd8
jart Remove bf16 luts
d6892c48
jart Make the GGML header look nicer
ce0442d7
jart Fix documentation
bc278c8a
jart Apply ggerganov's fixes for test-backend-ops
2741a997
jart Add BF16 code for new ggml_validate_row_data() function
632624e9
jart jart force pushed from 82aebcf0 to 632624e9 1 year ago
ggerganov ggerganov merged 38554160 into master 1 year ago
ddh0
arch-btw
jart
jart
teleprint-me
compilade
mofosyne mofosyne added Tensor Encoding Scheme
mofosyne mofosyne added Review Complexity : High

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone