issues
search
bytedance
/
lightseq
LightSeq: A High Performance Library for Sequence Processing and Generation
Other
3.22k
stars
329
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Interface optimization & decoder development
#379
hexisyztem
closed
2 years ago
0
[WIP] Optimize the interface design and complete the decoder layer development
#378
hexisyztem
closed
2 years ago
0
Crf layer
#377
neopro12
closed
2 years ago
0
RuntimeError: Unable to read HDF5 file from lightseq_bert_base_uncased.hdf5
#376
hurun
opened
2 years ago
1
What is the purpose of layers_new?
#375
melody-rain
closed
2 years ago
3
add debug log & canonical layer/node/tensor naming
#374
hexisyztem
closed
2 years ago
0
finish Encoder bw
#373
hexisyztem
closed
2 years ago
0
Remove LayerWeight, simplify the use of computational graphs
#372
hexisyztem
closed
2 years ago
0
TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
#371
zhuyan111
closed
2 years ago
2
Fix bart eval oom
#370
godweiyang
closed
2 years ago
0
Ls new arch - finish bert
#369
hexisyztem
closed
2 years ago
0
[WIP] New arch
#368
hexisyztem
closed
2 years ago
0
Add gradient communication quantization (GCQ)
#367
HandH1998
closed
2 years ago
4
New arch develop - encoder layer forward
#366
hexisyztem
closed
2 years ago
0
New arch encoder layer inference
#365
hexisyztem
closed
2 years ago
0
Crf
#364
neopro12
closed
2 years ago
0
test_launch_ffn_bias_bwd failed
#363
401qingkong
opened
2 years ago
2
support Google T5 and MT5 model
#362
anaivebird
closed
2 years ago
1
CUDA_HOME error? AssertionError: CUDA_HOME does not exist, unable to compile CUDA op(s)
#361
Mi-Peng
opened
2 years ago
4
add new kernels & operators
#360
hexisyztem
closed
2 years ago
0
fix compatibility with torch > 1.10
#359
Taka152
closed
2 years ago
0
[WIP] New arch develop - add new operators
#358
hexisyztem
closed
2 years ago
0
GPT2 推理怎么使用beam_search，并同时输出多个结果呢？
#357
wotulong
opened
2 years ago
1
请问，gpt2 转换成lightseq后，支持传入token_type_ids吗？
#356
yuanhuachao
opened
2 years ago
3
add lsflow
#355
neopro12
closed
2 years ago
0
Error when train and inference based on master code
#354
dongdaoking
opened
2 years ago
5
[WIP] New arch develop
#353
hexisyztem
closed
2 years ago
0
new_arch
#352
hexisyztem
closed
2 years ago
0
是否支持huggingface T5ForConditionalGenerartion的加速
#351
czq1999
closed
2 years ago
3
运行example中的ls_bert.py，报lsi.Bert报错
#350
yechong316
opened
2 years ago
2
CUDA ERROR
#349
ifromeast
opened
2 years ago
1
报错 CUBLAS_STATUS_NOT_SUPPORTED，不知道该怎么解决
#348
ImmortalZ
opened
2 years ago
2
请问量化功能什么时候发布正式tag
#347
BucherLi
opened
2 years ago
3
推理加速效果不错，cpu 300%下降到40%，速度提升 3倍
#346
ImmortalZ
opened
2 years ago
1
i want to inference this model by lightseq ，Which example should I refer to ?
#345
ImmortalZ
closed
2 years ago
1
support shard databin and hdfs for fairseq
#344
aachong
closed
2 years ago
0
i want to inference this model by lightseq
#343
ImmortalZ
closed
2 years ago
2
How to run the example of vit?
#342
Mi-Peng
opened
2 years ago
3
how to apply src_mask in LSTransformerEncoderLayer forward function?
#341
jxybb
opened
2 years ago
2
I donot know what does lightseq depend on?
#340
LiYijin
opened
2 years ago
1
AttributeError: module 'lightseq.inference' has no attribute 'Vit'
#339
chenjiashuo123
closed
2 years ago
1
How can I get multi results from once inference ?
#338
MowanHu
opened
2 years ago
5
Stop text feature for inference module
#337
hyunwoongko
opened
2 years ago
4
How to generate longer sentences
#336
koking0
closed
7 months ago
2
fix quant and torch layer bugs
#335
aachong
closed
2 years ago
0
support T5
#334
anaivebird
closed
2 years ago
0
fine-tune bart
#333
aachong
closed
2 years ago
0
Export bart
#332
aachong
closed
2 years ago
0
Export bart
#331
aachong
closed
2 years ago
0
Does lightseq support pipeline parallelism?
#330
dingjingzhen
opened
2 years ago
1
Previous
Next