issues
search
AI-Hypercomputer
/
maxtext
A simple, performant and scalable Jax LLM!
Apache License 2.0
1.47k
stars
275
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Test
#919
shralex
opened
16 hours ago
0
adding checkpointing storage options to base.yml
#918
rdyro
opened
16 hours ago
0
[DO NOT SUBMIT] Testing copybara again
#917
gobbleturk
closed
16 hours ago
0
Fixing image build error
#916
parambole
opened
18 hours ago
0
Small whitespace change test internal copybara migration
#915
gobbleturk
closed
21 hours ago
0
Training more than one epoch
#914
peregilk
opened
1 day ago
2
Use attn_mask_type of causal_padding for cudnn_flash_attention
#913
bvandermoon
opened
1 day ago
1
[Do not submit, testing copybara]
#912
gobbleturk
closed
21 hours ago
0
Support nsys profiler upload in all cases
#911
gobbleturk
opened
1 day ago
0
fix eval step in convergence test
#910
aireenmei
closed
17 hours ago
0
move gsutil copy to condtional to avoid breakages
#909
kocchop
closed
18 hours ago
6
Create collective microbenchmarks.
#908
qinyiyan
closed
2 days ago
1
Main merge
#907
kyle-google
opened
3 days ago
1
Give user option for activation type precision
#906
gobbleturk
closed
1 day ago
0
Making sure we run pylint only once, and run pyink in the same way.
#905
shralex
closed
4 days ago
0
Move maxtext docker images being built to artifact registry
#904
parambole
opened
4 days ago
0
Adds a new end-to-end test for Mistral 7b
#903
shralex
closed
2 days ago
0
Fix lint errors
#902
shralex
closed
6 days ago
0
Refactoring Maxtext build process with stable stack
#901
parambole
closed
1 day ago
0
Disable zarr3 when using single controller runtime
#900
shauryagup
closed
1 day ago
0
Fix lint errors.
#899
shralex
closed
6 days ago
0
Add configs for Attention Block Size tuning
#898
Obliviour
closed
6 days ago
0
Maxtext Offline serverless inference code
#897
vipannalla
closed
16 hours ago
2
[WIP] partial nnx impl
#896
rdyro
opened
1 week ago
0
Initialize jax distributed when checkpointing is enabled
#895
jonb377
opened
1 week ago
4
Add Llama 2 70B config on v5p
#894
raymondzouu
closed
5 days ago
0
Run code-style on changed files in pre-commit.
#893
shralex
closed
1 week ago
0
Run code-style on changed files in pre-commit
#892
shralex
closed
1 week ago
0
Add precision option
#891
RissyRan
closed
1 week ago
0
Docker prune in all github actions
#890
khatwanimohit
opened
1 week ago
1
Add GPT-3 175B v5p MLPerf 4.0 scripts
#889
anfals
closed
1 week ago
0
convert maxtext trained orbax checkpoint to HF checkpoint
#888
jwyang-google
closed
1 week ago
1
converted mlperf gpt3 ckpt starts with a worse loss
#887
gramesh-amd
opened
1 week ago
18
stage first axes mesh
#886
gobbleturk
closed
21 hours ago
0
Move pylint satements to the top of the file to conform with Google s…
#885
shralex
closed
1 week ago
0
Disable AOT activation offload test
#884
gobbleturk
closed
1 week ago
0
removing tensorflow_text for aarch64 compatiblity
#883
rdyro
opened
1 week ago
1
Support older checkpoint deletion and customized checkpoint sizes
#882
bernardhan33
closed
1 week ago
0
Add gpt-3 175B script for trillium
#881
raymondzouu
closed
1 week ago
0
Switch Expert axis to avoid unnecessary copy for layout change
#880
ZhiyuLi-goog
closed
1 week ago
2
Error loading mlperf gpt3 checkpoint after pax to maxtext conversion
#879
gramesh-amd
closed
1 week ago
14
Mask is being ignored when cudnn_flash_attention is used
#878
finbarrtimbers
opened
2 weeks ago
1
remove attention type from gemma2 model configs
#877
wenxindongwork
closed
1 week ago
2
Add eval to convergence test and log metrics
#876
aireenmei
closed
1 week ago
0
Cannot load the paxml gpt3 tokenizer
#875
gramesh-amd
closed
2 weeks ago
7
Additional Step to clean older docker images
#874
parambole
closed
2 weeks ago
0
[MoE][int8] add quantization to MoE dropped implementation
#873
ZhiyuLi-goog
closed
2 weeks ago
0
Fix mixtral8x7b model decoder references
#872
kyle-google
closed
2 weeks ago
1
Add block_until_ready operation before checkpoint saving operation.
#871
abhinavclemson
closed
2 weeks ago
0
clarify that Flax checkpoints are expected for Gemma
#870
nhira
closed
2 weeks ago
0
Next