total_score_history = []
best_ckpt_score = None
ckpt_file = os.path.join(models_folder, "summarizer_"+args.experiment+"_ckpt.bin")
ckpt_optimizer_file = os.path.join(models_folder, "summarizer_optimizer_"+args.experiment+"_ckpt.bin")
learning_rate = 2e-5
n_epochs = args.n_epochs
utils_hdf5.DoublePrint("printlog_summarizer_"+args.experiment+"_"+datetime.now().strftime("%Y-%m-%d")+".log", "a") ## << Wooh
if args.device == "cuda":
print("Training on GPU "+str(freer_gpu))
bert_tokenizer = utils_tokenizer.BERTCacheTokenizer()
print("---------------")
summarizer = GeneTransformer(max_output_length=args.max_output_length, device=args.device, tokenizer_type='gpt2', starter_model=args.model_start)
print("Summarizer loaded")
def collate_func(inps):
return [inp[0].decode() for inp in inps], [inp[1].decode() for inp in inps]
param_optimizer = list(summarizer.model.named_parameters())
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [
{'params': [p for n, p in param_optimizer if not any(nd in n for nd in no_decay)], 'weight_decay': 0.01},
{'params': [p for n, p in param_optimizer if any(nd in n for nd in no_decay)], 'weight_decay': 0.0}
]
logplot_file = os.path.join(args.log_folder, "gpt2_unsumm_"+args.experiment+".log")
logplot = LogPlot(logplot_file)
ckpt_file = os.path.join(models_folder,
"summarizer_" + args.experiment + "_ckpt.bin")
ckpt_optimizer_file = os.path.join(
models_folder, "summarizer_optimizer_" + args.experiment + "_ckpt.bin")
learning_rate = 2e-5
n_epochs = args.n_epochs
if args.device == "cuda":
print("Training on GPU " + str(freer_gpu))
bert_tokenizer = utils_tokenizer.BERTCacheTokenizer()
print("---------------")
summarizer = GeneTransformer(max_output_length=args.max_output_length,
device=args.device,
tokenizer_type='gpt2',
starter_model=summarizer_model_start)
print("Summarizer loaded")
def collate_func(inps):
if ".db" in args.dataset_file:
return [a['body'] for a in inps]
else:
return [inp[0].decode() for inp in inps]
param_optimizer = list(summarizer.model.named_parameters())
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params':
os.environ["CUDA_VISIBLE_DEVICES"] = "" + str(freer_gpu)
args.experiment += "_" + freer_gpu
learning_rate = 2e-5
n_epochs = args.n_epochs
utils_hdf5.DoublePrint("printlog_generator_" + args.experiment + "_" +
datetime.now().strftime("%Y-%m-%d") + ".log",
"a") ## << Wooh
bpe_model = ""
if args.tokenizer == "bpecap":
bpe_model = os.path.join(models_folder, "m.model")
model = GeneTransformer(tokenizer_type=args.tokenizer,
max_output_length=args.max_output_length,
device=args.device,
bpe_model=bpe_model)
if len(args.starter_model) > 0:
model.reload(os.path.join(models_folder, args.starter_model))
print("Model loaded")
def collate_func(inps):
return [inp[0] for inp in inps], [inp[1] for inp in inps]
dataset = utils_hdf5.HDF5Dataset(args.dataset_file, collection_name="name")
N = len(dataset)
N_dev = 500