|
What should I do if I want to use model from DeepSpeed
|
|
5
|
1601
|
April 6, 2024
|
|
[Maybe Bug] When using EarlyStopping Callbacks with Seq2SeqTraininer, training didn't stop
|
|
3
|
1466
|
April 4, 2024
|
|
ncclSystemError: System call (e.g. socket, malloc) or external library call failed or device error
|
|
0
|
1076
|
March 30, 2024
|
|
Deepspeed zero-2 cpu offloading killing process = -9 error
|
|
1
|
1633
|
March 17, 2024
|
|
Conceptual question: Early loading of the model defeats the purpose of deepspeed!
|
|
0
|
156
|
March 14, 2024
|
|
Struggle with finetuneing flan-t5-xxl using deepspeed
|
|
3
|
833
|
March 12, 2024
|
|
Deepspeed inference stage 3 + quantization
|
|
0
|
907
|
March 8, 2024
|
|
Saving checkpoint is too slow with deepspeed
|
|
5
|
2525
|
March 6, 2024
|
|
Deepspeed trainer and custom loss weights
|
|
1
|
540
|
February 28, 2024
|
|
How can I use Inference API with my model?
|
|
0
|
146
|
February 24, 2024
|
|
Finetune LLM with DeepSpeed
|
|
2
|
4992
|
February 22, 2024
|
|
DeepSpeed integration for HuggingFace Seq2SeqTrainingArguments
|
|
0
|
1371
|
February 22, 2024
|
|
It says that `bfloat16.enabled` without `auto' needed to be specified when training T5, is anyone aware of how to solve that?
|
|
0
|
243
|
February 20, 2024
|
|
Exact difference between Transformers' and Accelerate's DeepSpeed integrations?
|
|
5
|
740
|
February 13, 2024
|
|
How to use GPU when using transformers.AutoModel
|
|
0
|
1535
|
February 3, 2024
|
|
Multi GPU training - Model parallelism
|
|
1
|
1800
|
February 2, 2024
|
|
More processes than GPUs with DeepSpeed launcher
|
|
0
|
219
|
January 25, 2024
|
|
Rewrite trainer's save_model method get unexpected pytorch_model.bin file
|
|
0
|
377
|
January 8, 2024
|
|
Model (Pipeline) Parallelism in SLURM cluster
|
|
0
|
234
|
January 6, 2024
|
|
Mixtral bad FP16 performance
|
|
0
|
510
|
January 3, 2024
|
|
Deepspeed script launcher vs accelerate script launcher for TRL
|
|
0
|
354
|
December 25, 2023
|
|
Best practice to run DeepSpeed
|
|
2
|
1526
|
December 25, 2023
|
|
Infrence time increase when using multi-GPU
|
|
1
|
877
|
November 28, 2023
|
|
Resume_from_checkpoint does not configure learning rate scheduler correctly
|
|
3
|
851
|
November 28, 2023
|
|
What does LoRA do to model by default?
|
|
0
|
528
|
November 21, 2023
|
|
The same hyperparameters with deepspeed is worse than without deepseepd
|
|
2
|
437
|
November 13, 2023
|
|
Deepspeed stage 3 partition
|
|
0
|
575
|
October 31, 2023
|
|
Unable to train model (Loss is 0.000000)
|
|
2
|
1059
|
October 17, 2023
|
|
Pix2struct based model ddp code conversion
|
|
1
|
310
|
October 11, 2023
|
|
Speed up beam search for item generation
|
|
1
|
913
|
October 4, 2023
|