Default parameters for 8G VRAM

main
BlinkDL 4 years ago
parent 6aefe59c3d
commit 0b6aec3da6

@ -16,7 +16,7 @@ logger = logging.getLogger(__name__)
########################################################################################################
T_MAX = 1024 # increase this if your ctx_len > 1024
B_GROUP_FORWARD = 8 # set to 8 for best performance
B_GROUP_FORWARD = 4 # set to 8 for best performance
B_GROUP_BACKWARD = 2 # set to 2 for best performance
timex_cuda = load(name="timex", sources=["cuda/timex_op.cpp", "cuda/timex_cuda.cu"],

@ -34,14 +34,14 @@ model_type = 'RWKV'
# ===> batch_size must be divisible by B_GROUP_FORWARD and B_GROUP_BACKWARD in model.py
# For example, if your batch_size = 20, you can set B_GROUP_FORWARD = 4, B_GROUP_BACKWARD = 2
# If you see "CUDA out of memory", reduce it. Use GPU-Z to find the highest value for your VRAM.
batch_size = 40
batch_size = 12
### Step 4: set learning rate, training 'epochs' #######################################################
lr_init = 6e-4
lr_final = 1e-5
# the 'epoch' here is very short and of fixed length (ctx_len * epoch_length_fixed tokens)
n_epoch = 1000
n_epoch = 500
# 0 = never, 1 = every 'epoch', 2 = every two 'epoch', etc.
epoch_save_frequency = 30
epoch_save_path = 'trained-'
@ -124,8 +124,8 @@ if __name__ == '__main__':
model = GPT(GPTConfig(train_dataset.vocab_size, train_dataset.ctx_len, model_type=model_type,
n_layer=n_layer, n_embd=n_embd)).cuda()
# # load a trained model. remember to change random seed
# m2 = torch.load('trained-10000.pth')
# # # load a trained model. remember to change random seed
# m2 = torch.load('trained-61.pth')
# model.load_state_dict(m2)
print('model', model_type, 'epoch', n_epoch, 'batchsz', batch_size, 'betas',
@ -137,5 +137,5 @@ if __name__ == '__main__':
trainer.train()
torch.save(model, 'trained-' + str(n_epoch) + trainer.get_run_name() +
torch.save(model, 'trained-' + str(n_epoch) + '-' + trainer.get_run_name() +
'-' + datetime.datetime.today().strftime('%Y-%m-%d-%H-%M-%S') + '.pth')

Loading…
Cancel
Save