Go
Home
Pricing
FAQ
Install
Home
Pricing
FAQ
Install
Login
via GitHub
huggingface/accelerate
Pull Requests
Commits
muellerzr-msamp-ds-fsdp
3d-parallelism
argparse
better-err
big_api
check-docs
check-for-nccl
composable-tp
context-parallel
context-parallel-experiments
context-parallel-flex-attn
cp-dataloader
cp-pc
dataloader-log
debug-tests
deepspeed-inference
deepspeed-version
device_map_xla_support
disable-seedale-rs
enable-dash
feat/async-checkpointing
feat-decorator-to-purge-modified-accelerate-env-vars
fix
fix-compile-regions
fix-deepspeed-autobs
fix-dispatch-model-tied-params-memory
fix-fp8
fix-generate
fix-grad-norm
fix-pjrt_device
fix-prod
fix-warnings
fork-tester
fp8-gradient-checkpointing
fp8-stuff
fsdp2-tp
fully-remove-accelerate-config
grad-acc-optimizer-fixes
grad-accum-test
import-util
llama-to-mistral
load-model-across-devices
low-bit-fsdp2
main
make-version-tests-better
mishig25-patch-1
mishig25-patch-2
mixed-precision-experiments
ms-amp
muellerzr-ds-debugging
muellerzr-fix-1.0
muellerzr-fp8-deepspeed-support-v2
muellerzr-msamp-ds-fsdp
muellerzr-nightly-fixings
muellerzr-stateful-dl
new-instance-type
nouamane/context-parallel
parallelism-config
pin-ruff
pip-uv
pippy-duplicates
pippy-integration
reaction-based-runs
release-v0.6.1
release-v0.6.2
revert-3671
revert-fsdp-improv
revert-pr
rm-112
runner
safetensors-default
slack-reporter
speedup-docker
test-data
test-deepspeed-unpin
torch-22
trainer-tests
transformers-nd-parallel
ulysses-sp
unfreeze-4090
use-partialstate
uv-take2
v0.7-release
v0.12-release
v0.13-release
v0.14-release
v0.15-release
v0.16-release
v0.17-release
v0.18-release
v0.19-release
v0.20-release
v0.21-release
v0.22-release
v0.23-release
v0.24-release
v0.25.0-release
v0.26.0-release
v0.26.1-release
v0.27.0-release
v0.28.0-release
v0.29.0-release
v0.30.0-release
v0.31.0-release
v0.32.0-release
v0.33.0-release
v0.34.0-release
v1.0.0-release
v1.1.0-release
v1.2.0-release
v1.3.0-release
v1.4.0-release
v1.5.0-release
v1.6.0-release
v1.7.0-release
v1.8.0-release
v1.9.0-release
v1.10.0-release
v1.11.0-release
v1.12.0-release
wip-from-pretrained
xla-gpu-runners
Format
muellerzr
committed
1 year ago
e5f6d570
Add msamp init args to deal with weight qtype defaults
muellerzr
committed
1 year ago
854e44a0
Clean and fin
muellerzr
committed
1 year ago
4a37e5f0
checkpoint
muellerzr
committed
1 year ago
21569352
Continue
muellerzr
committed
1 year ago
e7979c30
Bookmark
muellerzr
committed
1 year ago
019d14da
Working!
muellerzr
committed
1 year ago
7db3368b
Snapshot: Works
muellerzr
committed
1 year ago
41ce97ca
Freeze: working version
muellerzr
committed
1 year ago
f3fe0d07
Working base fsdp version with only MSAMP
muellerzr
committed
1 year ago
8da59ac9
Working MS-AMP with non-distributed and DDP
muellerzr
committed
1 year ago
bb4061c7
Fixed MS-AMP implementation
muellerzr
committed
1 year ago
64552d49
Off, but we're training now at least
muellerzr
committed
1 year ago
8ed5816d
Working MS-AMP implementation + some fixes for accelerate
muellerzr
committed
1 year ago
569c23da
destroy process group in `end_training` (#3012)
SunMarc
committed
1 year ago
Verified
589fddd3
Wrong import check for TE (#3016)
muellerzr
committed
1 year ago
Verified
99c69aaf
fix default value for rank size in cpu threads_per_process assignment logic (#3009)
rbrugaro
committed
1 year ago
Verified
00785cd9
Enable FSDP & Deepspeed + FP8 (#2983)
muellerzr
committed
1 year ago
Verified
a452327e
Fix `find_tied_params` for models with shared layers (#2986)
qubvel
committed
1 year ago
Verified
851cf343
update version to 0.34.dev0 (#3007)
SunMarc
committed
1 year ago
Verified
cd5698bb
Add small util to enable FSDP offloading quickly (#3006)
muellerzr
committed
1 year ago
Verified
90d50239
Make env variables optional for FSDP (#2998)
muellerzr
committed
1 year ago
Verified
3bde6156
Fix deepspeed tests (#3003)
muellerzr
committed
1 year ago
Verified
dc3b5ad8
clear memory after offload (#2994)
SunMarc
committed
1 year ago
Verified
12a5befd
Support skip_first_batches for XLA (#2966)
yitongh
committed
1 year ago
Verified
79ca85c2
Fix typo on warning str: "meta device device" -> "meta device" (#2997)
HeAndres
committed
1 year ago
Verified
13d93c4f
Explicit check for `step` when loading the state (#2992)
muellerzr
committed
1 year ago
Verified
d982751a
Fix gated test (#2993)
muellerzr
committed
1 year ago
Verified
95edc68c
Fix bug of clip_grad_norm_ for xla fsdp (#2941)
append-only
committed
1 year ago
Verified
288accc0
remove .md to allow proper linking (#2977)
nbroad1881
committed
1 year ago
Verified
83b06101
Older