ruotianluo / ImageCaptioning.pytorch

I decide to sync up this repo and self-critical.pytorch. (The old master is in old master branch for archive)
MIT License
1.44k stars 412 forks source link

测试时报错 #158

Open zhouhaonan011 opened 2 years ago

zhouhaonan011 commented 2 years ago

File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 240, in eval_split lang_stats = language_eval(dataset, predictions, n_predictions, eval_kwargs, split) File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 91, in language_eval meanperplexity = sum([['perplexity'] for _ in preds_filt]) / len(preds_filt) ZeroDivisionError: division by zero

至今还是没有想明白这是什么原因,我也不是重复测试啊,第一次就这样了

ruotianluo commented 2 years ago

开了force也没有用吗

Ruotian Luo

On Jun 13, 2022, at 6:41 PM, chuanxilu @.***> wrote:

 File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 240, in eval_split lang_stats = language_eval(dataset, predictions, n_predictions, eval_kwargs, split) File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 91, in language_eval meanperplexity = sum([['perplexity'] for _ in preds_filt]) / len(preds_filt) ZeroDivisionError: division by zero

至今还是没有想明白这是什么原因,我也不是重复测试啊,第一次就这样了

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.

zhouhaonan011 commented 2 years ago

开了force也没有用吗 Ruotian Luo On Jun 13, 2022, at 6:41 PM, chuanxilu @.***> wrote:  File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 240, in eval_split lang_stats = language_eval(dataset, predictions, n_predictions, eval_kwargs, split) File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 91, in language_eval meanperplexity = sum([['perplexity'] for _ in preds_filt]) / len(preds_filt) ZeroDivisionError: division by zero 至今还是没有想明白这是什么原因,我也不是重复测试啊,第一次就这样了 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.

是的,加了force,也是这样

ruotianluo commented 2 years ago

你还有改其他东西吗?

Ruotian Luo

On Jun 13, 2022, at 7:40 PM, chuanxilu @.***> wrote:

 开了force也没有用吗 Ruotian Luo … On Jun 13, 2022, at 6:41 PM, chuanxilu @.***> wrote:  File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 240, in eval_split lang_stats = language_eval(dataset, predictions, n_predictions, eval_kwargs, split) File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 91, in language_eval meanperplexity = sum([['perplexity'] for _ in preds_filt]) / len(preds_filt) ZeroDivisionError: division by zero 至今还是没有想明白这是什么原因,我也不是重复测试啊,第一次就这样了 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.

是的,加了force,也是这样

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.

zhouhaonan011 commented 2 years ago

你还有改其他东西吗? Ruotian Luo On Jun 13, 2022, at 7:40 PM, chuanxilu @.> wrote:  开了force也没有用吗 Ruotian Luo … On Jun 13, 2022, at 6:41 PM, chuanxilu @.> wrote:  File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 240, in eval_split lang_stats = language_eval(dataset, predictions, n_predictions, eval_kwargs, split) File "/home/ImageCaptioning.pytorch-master/eval_utils.py", line 91, in language_eval meanperplexity = sum([['perplexity'] for _ in preds_filt]) / len(preds_filt) ZeroDivisionError: division by zero 至今还是没有想明白这是什么原因,我也不是重复测试啊,第一次就这样了 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread. 是的,加了force,也是这样 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.

刚才按照您以前提供的方法,测试的时候没有问题了。 但是出现了新的问题:当我把caption model从 newfc 替换成show_tell, show_attend_tell 这些模型进行训练的时候,会报错: Command line argument and saved model disagree on caption_model 可是我并没有加载已有的模型,这是什么情况,真的给我搞晕了。我怀疑是没有用新的文件夹来保存模型的原因进行训练,我就换了个文件夹来存放新的show_attend_tell模型。但是还是无法正常训练。

ruotianluo commented 2 years ago

嗯,这两个模型可能就是太老了,我后来就没支持了

zhouhaonan011 commented 2 years ago

嗯,这两个模型可能就是太老了,我后来就没支持了

好的,感谢您的耐心解答