issues
search
microsoft
/
tutel
Tutel MoE: An Optimized Mixture-of-Experts Implementation
MIT License
694
stars
84
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
removed logit_scale without device casting
#191
Harsh-Sensei
closed
1 year ago
1
RuntimeError: No such operator tutel_ops::cumsum
#190
sharkdrop
opened
1 year ago
10
[installation errors] fatal error: nccl.h: No such file or directory
#189
qianyuzqy
opened
1 year ago
1
fix typos and old pytorch compatibility
#188
ghostplant
closed
1 year ago
0
Multi-nodes training is much more slower than single node
#187
YingqingHe
opened
1 year ago
1
New Tutel checkpoint loading is incompatible with old models
#186
jinga-lala
closed
1 year ago
7
NCCL Asynchronous update timeout crash with Tutel MoE
#185
jinga-lala
opened
1 year ago
5
extend parallel_type for adaptive:n
#184
ghostplant
closed
2 years ago
0
extend parallel_type to use dp without a2a
#183
ghostplant
closed
2 years ago
0
My code seems to hang when skip_remainder_batch=False.
#182
Fragile-azalea
opened
2 years ago
7
support tutel.checkpoint.* for issue #177
#181
ghostplant
closed
2 years ago
0
Cannot import JIT optimized kernels. Did you forget to install Custom Kernel Extension?
#180
Alex-Songs
opened
2 years ago
1
Pretrained MoE model
#179
Luodian
opened
2 years ago
2
Example on saving experts to one model when using distributed training
#178
Luodian
opened
2 years ago
2
Error when doing deepcopy of the model
#177
yzxing87
opened
2 years ago
5
add tensor save/load in numpy format
#176
ghostplant
closed
2 years ago
0
[installation errors] fatal error: nccl.h: No such file or directory
#175
Luodian
closed
2 years ago
1
a bunch of fixes for #167 and #173
#174
ghostplant
closed
2 years ago
0
Is simple_all_reduce also required for capacity_factor > 0 cases?
#173
Fragile-azalea
closed
2 years ago
6
The output of nccl_all_to_all_scatter_async may be incomplete when num_local_experts>1.
#172
Fragile-azalea
closed
2 years ago
11
Cannot Import JIT optimized kernels?
#171
Luodian
closed
2 years ago
11
handle Windows Pytorch compatibility
#170
ghostplant
closed
2 years ago
0
how can I install this pack on conda environment??
#169
Lurnco
opened
2 years ago
11
typo fix
#168
ghostplant
closed
2 years ago
0
Error in load_importance_loss
#167
Luodian
opened
2 years ago
7
update TutelDistributedOptimizer
#166
zeliu98
opened
2 years ago
0
refine fairseq_moe configuration
#165
ghostplant
closed
2 years ago
0
allow CUDA_HOME to specify CUDA SDK location
#164
ghostplant
closed
2 years ago
0
Cannot compile tutel kernels and got runtime error
#163
hyhuang00
closed
2 years ago
10
Add Feature - Port overlapping from v0.1.x and support per-layer overlapping degree
#162
yzygitzh
closed
2 years ago
0
bp of shared parameters and experts
#161
a157801
opened
2 years ago
7
100x slower when using 4nodes than 1node to run the helloworld_ddp example
#160
a157801
closed
2 years ago
12
update test case in pipeline
#159
ghostplant
closed
2 years ago
0
update README.md
#158
ghostplant
closed
2 years ago
0
add error reasons for installation
#157
ghostplant
closed
2 years ago
0
AttributeError: module 'tutel_custom_kernel' has no attribute 'inject_source'
#156
s-kodge
closed
2 years ago
3
add cosine router; add load loss and importance loss
#155
zeliu98
closed
2 years ago
0
Add example for fairseq moe with tutel support
#154
EricWangCN
closed
2 years ago
0
add examples: tutel.examples.helloworld_ddp_tutel
#153
ghostplant
closed
2 years ago
0
add examples: tutel.examples.helloworld_tutel_ddp
#152
ghostplant
closed
2 years ago
0
add simple patch for Fairseq using MoE
#151
ghostplant
closed
2 years ago
0
move `fp32_gate` checking from moe_layer to top
#150
ghostplant
closed
2 years ago
0
allow moe.moe_layer to use custom expert
#149
ghostplant
closed
2 years ago
0
add net.barrier()
#148
ghostplant
closed
2 years ago
0
refine argument list in custom gate module
#147
ghostplant
closed
2 years ago
0
add examples: tutel.examples.moe_cifar10
#146
ghostplant
closed
2 years ago
0
support example: tutel.examples.moe_cifar10
#145
ghostplant
closed
2 years ago
0
distinguish cpu/gpu switching
#144
ghostplant
closed
2 years ago
0
fix gradient update in MNIST example
#143
ghostplant
closed
2 years ago
0
add example: tutel.examples.moe_mnist
#142
ghostplant
closed
2 years ago
0
Previous
Next