issues
search
apple
/
axlearn
An Extensible Deep Learning Library
Apache License 2.0
1.86k
stars
259
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Adds support for shared modules in auto-context handling.
#804
markblee
opened
5 hours ago
0
Add health check before trainer launch
#803
hanzhi713
opened
1 day ago
0
Implements DecoderMixin as a DecodingLayer.
#802
markblee
closed
1 day ago
0
Correct CAUSAL padding in conv when stride > 1.
#801
ds-hwang
closed
2 days ago
2
Remove `jax_enable_memories` config
#800
hanzhi713
closed
2 days ago
0
Make some convolution functions public.
#799
ds-hwang
closed
2 days ago
1
Skip GPU test for unsupported unequal length case
#798
qdavid1
closed
2 days ago
0
Introduce @nowrap annotaion.
#797
ds-hwang
closed
3 days ago
1
Adds optional autotune optimization to input_tf_data.sample_from_datasets.
#796
tgunter
closed
2 days ago
0
Test a switch from CircleCI to Github Actions
#795
madrob
opened
4 days ago
0
Bump up circle CI container size for build-and-test.
#794
kelvin-zou
closed
4 days ago
0
Key and value input for MHA prefill and extend
#793
qdavid1
closed
4 days ago
0
[DO NOT CHECK IN, TEST ONLY]Test CI OOM issue
#792
kelvin-zou
closed
4 days ago
0
Make RepeatedConformerLayer's repeat field configurable.
#791
ds-hwang
closed
5 days ago
1
Revert "Speed up Axlearn CI"
#790
kelvin-zou
closed
4 days ago
2
Weight only offload
#789
hanzhi713
closed
2 days ago
0
Remove init workaround
#788
hanzhi713
closed
5 days ago
0
Add multislice XLA flag to improve tf summary metric induced latency.
#787
tgunter
closed
5 days ago
0
Add TPU Monitoring for faster hang detection
#786
kelvin-zou
closed
4 days ago
0
improve groupnorm
#785
berlino
closed
4 days ago
0
Adds grain input dispatch.
#784
markblee
closed
5 days ago
0
Add Goodput & Badput recording and monitoring support.
#783
dipannita08
opened
1 week ago
1
tokens_per_batch fixed to take into account DP and micro-batch accumulation steps
#782
amithrm
closed
1 week ago
0
Simplifies host_to_global_device_array.
#781
markblee
closed
5 days ago
0
Support Causal Convolution.
#780
ds-hwang
closed
5 days ago
2
Disambiguate source_len / target_len
#779
changlan
closed
1 week ago
1
load llama v3 weights into fuji
#778
sychen52
closed
5 days ago
1
Reshard the train state after restoring from builder
#777
changlan
closed
1 week ago
0
Fix inconsistent paddings in conv layer.
#776
ds-hwang
closed
1 week ago
1
Use fewer bytes for the NumpyMask
#775
changlan
closed
1 week ago
0
Significantly speedup ConfigBase
#774
soundway
closed
1 week ago
0
fsdp=16 model=16 gbs=16 should work on 256 chips
#773
samos123
opened
1 week ago
9
Support more types of groupnorm
#772
berlino
closed
5 days ago
2
restore test_apply_paddings_check runtime_checks test
#771
mattjj
opened
1 week ago
2
Add implicit-dirs to default gcsfuse settings.
#770
RsEnts
closed
1 week ago
0
Add neuron attention with tests
#769
lipovsek-aws
closed
1 week ago
0
Checkpointing with grain.
#768
markblee
closed
1 week ago
0
Configure output-uploader as a sidecar container
#767
Ethanlm
closed
1 week ago
0
add Fuji v3 405b model config
#766
samos123
opened
1 week ago
5
Adds XLA SDC check configuration to the trainer.
#765
tgunter
closed
1 week ago
0
audio pillar starts to use einops
#764
ds-hwang
closed
1 week ago
0
Expose NODE_IP to container env
#763
Ethanlm
closed
1 week ago
0
cannot install on Apple Silicon
#762
chunyang-wen
opened
1 week ago
2
Adds support for unequal lengths for query vs. key-value to attention logic
#761
qdavid1
closed
2 weeks ago
0
Adds grain lm eval inputs.
#760
markblee
closed
1 week ago
0
Upgrade axlearn to jax 0.4.33
#759
matthew-e-hopkins
closed
1 week ago
0
Sparse Sliding Window Attention
#758
changlan
closed
1 week ago
0
Optimizer offloading draft and experiments
#757
hanzhi713
opened
2 weeks ago
0
Alternative method of submitting jobs to DF Runner
#756
remylouisew
opened
2 weeks ago
2
Add host mout
#755
hanzhi713
closed
2 weeks ago
1
Next