thtrieu / darkflow

Translate darknet to tensorflow. Load trained weights, retrain/fine-tune using tensorflow, export constant graph def to mobile devices
GNU General Public License v3.0
6.13k stars 2.08k forks source link

How checkpoint files are created ? #1132

Open anushabhura opened 4 years ago

anushabhura commented 4 years ago

I am training my own model .

step 124 - loss 60.73698425292969 - moving ave loss 64.46739123826117 Finish 124 epoch(es) step 125 - loss 60.53150177001953 - moving ave loss 64.073802291437 Traceback (most recent call last): File "flow", line 6, in cliHandler(sys.argv) File "C:\Users\intel\darkflow\cli.py", line 33, in cliHandler print('Enter training ...'); tfnet.train() File "C:\Users\intel\darkflow\net\flow.py", line 65, in train if not ckpt: _save_ckpt(self, *args) File "C:\Users\intel\darkflow\net\flow.py", line 21, in _save_ckpt with open(profile, 'wb') as profile_ckpt: FileNotFoundError: [Errno 2] No such file or directory: './ckpt/cfg/tiny-yolo-vo c-4c-125.profile'

For me, it's working till 125 epochs and then this error is coming. How checkpoint file is created through program? I didn't get it