You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I use the following configuration for training, but when training epoch21, the accuracy has been 60.00 to 60.01 does not rise, Is this because I did not use 8*512 batchsize for training? could you please give me some advice, thanks!!!!
Hi, I use the following configuration for training, but when training epoch21, the accuracy has been 60.00 to 60.01 does not rise, Is this because I did not use 8*512 batchsize for training? could you please give me some advice, thanks!!!!
export NNODES=1
export GPUS_PER_NODE=2
export WANDB__SERVICE_WAIT=60
export CUDA_VISIBLE_DEVICES=4,5
DISTRIBUTED_ARGS="--nproc_per_node $GPUS_PER_NODE --nnodes $NNODES "
torchrun $DISTRIBUTED_ARGS src/training/main2.py
--save-frequency 1
--report-to wandb
--train-data /home/gg/gg/MQBench-main/test/model/e1/split_2tar
--dataset-type webdataset
--imagenet-val ./ImageNet
--warmup 2000
--batch-size 2048
--epochs 25
--workers 16
--model TinyCLIP-ViT-39M-16-Text-19M
--name exp_name2
--seed 0
--local-loss
--grad-checkpointing
--output ./outputs/c
--lr 0.0001
--gather-with-grad
--pretrained-image-file ViT-B-16@openai
--pretrained-text-file ViT-B-16@openai
--distillation-teacher ViT-B-32@laion2b_e16
--norm_gradient_clip 5
--train-num-samples 15000000
--logit-scale 50
The text was updated successfully, but these errors were encountered: