Could you share the trained model parameters?

[Yc174]

Could you share the trained model parameters?On coco or voc? @xmyqsh

[chl916185]

Have you finished the training, I have a problem loading the training model (Resnet50.npy),Could you share the Resnet50.npy?

[Yc174]

I have trained 100000 steps,but the results are not good.Here are the links of Resnet.npy @chl916185

[BestSongEver]

Hi, I have trained the model with my own data , but i don't know how to run a test to get the result.(To get the mAP of my task). Could u please tell my how to run a test? Thank u so so so so so so much. @Yc174

[chl916185]

assign pretrain model weights to res3d_branch2b assign pretrain model weights to res3d_branch2c assign pretrain model weights to res3d_branch2a cudaCheckError() failed in ROIPoolForward: invalid device function F tensorflow/stream_executor/cuda/cuda_driver.cc:413] Check failed: CUDA_SUCCESS == dynloa d::cuCtxSetCurrent(cuda_context->context()) (0 vs. 4) E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:656] failed to record completion ev ent; therefore, failed to create inter-stream dependency I tensorflow/stream_executor/stream.cc:3775] stream 0x6bd8970 did not memcpy device-to-hos t; source: 0x232c475a00 E tensorflow/stream_executor/cuda/cuda_driver.cc:1166] could not wait stream on event: CUD A_ERROR_DEINITIALIZED I tensorflow/stream_executor/stream.cc:3788] stream 0x6bd7db0 did not memcpy host-to-devic e; source: 0x7f7f9444c000 E tensorflow/stream_executor/cuda/cuda_event.cc:49] Error polling for event status: failed to query event: CUDA_ERROR_DEINITIALIZED F tensorflow/core/common_runtime/gpu/gpu_event_mgr.cc:198] Unexpected Event status: 1 I tensorflow/stream_executor/stream.cc:1338] stream 0x6bd8970 did not wait for stream: 0x6 bd73a0 I tensorflow/stream_executor/stream.cc:3775] stream 0x6bd8970 did not memcpy device-to-hos t; source: 0x23797d0c00 ./experiments/scripts/FPN_end2end.sh: line 54: 3942 Aborted (core dumped) python ./faster_rcnn/train_net.py --gpu ${GPU_ID} --weights data/pretrain_model/Resnet50. npy --imdb ${TRAIN_IMDB} --iters ${ITERS} --cfg ${CFG} --network FPN_train --restore 0 ${E XTRA_ARGS}

When I loaded Resnet50.npy , I met the above question, why? @Yc174 @BestSongEver

[BestSongEver]

Sorry, I never met the problem. U can google it. It seems like a problem with ur CUDA. @chl916185

[xmyqsh]

I'm training... Currently on voc

[iFighting]

how about the result?