SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

fixed the bug raised by "tmp_eval_loss += tmp_eval_loss.item()" when parallelly using multi-gpu.

Browse Source
This commit is contained in:
focox@qq.com
2019-10-23 20:27:13 +08:00
parent ef1b8b2ae5
commit bd847ce7d7
1 changed files with 3 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
examples/run_ner.py
View File

@@ -210,6 +210,9 @@ def evaluate(args, model, tokenizer, labels, pad_token_label_id, mode, prefix=""
outputs = model(**inputs) outputs = model(**inputs)
tmp_eval_loss, logits = outputs[:2] tmp_eval_loss, logits = outputs[:2]
if args.n_gpu > 1:
tmp_eval_loss = tmp_eval_loss.mean() # mean() to average on multi-gpu parallel evaluating
eval_loss += tmp_eval_loss.item() eval_loss += tmp_eval_loss.item()
nb_eval_steps += 1 nb_eval_steps += 1
if preds is None: if preds is None: