missing models_baseline.mlp

[daebong]

I have the following error while running python3 run_evaluate.py --posenet_name 'videopose' --keypoints gt --evaluate 'checkpoint/adaptpose/videopose/gt/3dhp/ckpt_best_dhp_p1.pth.tar' --dataset_target 3dhp --keypoints_target 'gt' --pad 13 --pretrain_path 'checkpoint/pretrain_baseline/videopose/gt/3dhp/ckpt_best.pth.tar' :

ModuleNotFoundError: No module named 'models_baseline.mlp'

I also had several other errors, and I wonder if you actually tried this package on a new environment :)

[YC-Yuan]

I meet the same error. The following might help, but I haven't tried it yet. The models_baseline package is in the PoseAug repository click here. I will reply again if the package works.

[YC-Yuan]

The missing module _modelsbaseline package is not the only missing dependency. Check VideoPose3D and AdaptPose to find all. I succeed in running _runevaluate.py & _runadaptpose.py, and get scores as in the paper.

[LLLYLong]

The missing module _modelsbaseline package is not the only missing dependency. Check VideoPose3D and AdaptPose to find all. I succeed in running _runevaluate.py & _runadaptpose.py, and get scores as in the paper.

Hi, I would like to ask, training on S1 at human 3.6m and evaluating at S5,S6,S7,S8, is S5,S6,S7,S8 viewed as a whole when training, or are they trained separately for S1, S5 so individually

[mgholamikn]

Hi, Yes, s5,s6,.. viewed as a whole when training. I didn’t train separately for s5,s6,…. Btw, evaluation is performed on the test set which is s9 and s11 similar to the previous work in the results. S1 is used as the source and s5,s6,s7,s8 used as target.

On Fri, Aug 30, 2024 at 7:31 PM LLLYLong @.***> wrote:

The missing module models_baseline package is not the only missing dependency. Check VideoPose3D https://github.com/facebookresearch/VideoPose3D and AdaptPose https://github.com/jfzhang95/PoseAug to find all. I succeed in running run_evaluate.py & run_adaptpose.py, and get scores as in the paper.

Hi, I would like to ask, training on S1 at human 3.6m and evaluating at S5,S6,S7,S8, is S5,S6,S7,S8 viewed as a whole when training, or are they trained separately for S1, S5 so individually

— Reply to this email directly, view it on GitHub https://github.com/mgholamikn/AdaptPose/issues/4#issuecomment-2322736336, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANJS3Z4XJFSWMT6V42HANU3ZUETGNAVCNFSM6AAAAABNNN5ZLOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRSG4ZTMMZTGY . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[LLLYLong]

Hi, Yes, s5,s6,.. viewed as a whole when training. I didn’t train separately for s5,s6,…. Btw, evaluation is performed on the test set which is s9 and s11 similar to the previous work in the results. S1 is used as the source and s5,s6,s7,s8 used as target. … On Fri, Aug 30, 2024 at 7:31 PM LLLYLong @.> wrote: The missing module models_baseline package is not the only missing dependency. Check VideoPose3D https://github.com/facebookresearch/VideoPose3D and AdaptPose https://github.com/jfzhang95/PoseAug to find all. I succeed in running run_evaluate.py & run_adaptpose.py, and get scores as in the paper. Hi, I would like to ask, training on S1 at human 3.6m and evaluating at S5,S6,S7,S8, is S5,S6,S7,S8 viewed as a whole when training, or are they trained separately for S1, S5 so individually — Reply to this email directly, view it on GitHub <#4 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANJS3Z4XJFSWMT6V42HANU3ZUETGNAVCNFSM6AAAAABNNN5ZLOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRSG4ZTMMZTGY . You are receiving this because you are subscribed to this thread.Message ID: @.>

@mgholamikn Thank you very much for your reply. That is, the training uses S1 as the source domain and S5,S6,S7,S8 as the target domain. The evaluation is testing the trained model on S9, S11 and not on S5, S6 etc.

[mgholamikn]

Yes, that’s correct

On Fri, Aug 30, 2024 at 10:59 PM LLLYLong @.***> wrote:

Hi, Yes, s5,s6,.. viewed as a whole when training. I didn’t train separately for s5,s6,…. Btw, evaluation is performed on the test set which is s9 and s11 similar to the previous work in the results. S1 is used as the source and s5,s6,s7,s8 used as target. … <#m-6243199253333627963> On Fri, Aug 30, 2024 at 7:31 PM LLLYLong @.> wrote: The missing module models_baseline package is not the only missing dependency. Check VideoPose3D https://github.com/facebookresearch/VideoPose3D https://github.com/facebookresearch/VideoPose3D and AdaptPose https://github.com/jfzhang95/PoseAug https://github.com/jfzhang95/PoseAug to find all. I succeed in running run_evaluate.py & run_adaptpose.py, and get scores as in the paper. Hi, I would like to ask, training on S1 at human 3.6m and evaluating at S5,S6,S7,S8, is S5,S6,S7,S8 viewed as a whole when training, or are they trained separately for S1, S5 so individually — Reply to this email directly, view it on GitHub <#4 (comment) https://github.com/mgholamikn/AdaptPose/issues/4#issuecomment-2322736336>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANJS3Z4XJFSWMT6V42HANU3ZUETGNAVCNFSM6AAAAABNNN5ZLOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRSG4ZTMMZTGY https://github.com/notifications/unsubscribe-auth/ANJS3Z4XJFSWMT6V42HANU3ZUETGNAVCNFSM6AAAAABNNN5ZLOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRSG4ZTMMZTGY . You are receiving this because you are subscribed to this thread.Message ID: @.>

@mgholamikn https://github.com/mgholamikn Thank you very much for your reply. That is, the training uses S1 as the source domain and S5,S6,S7,S8 as the target domain. The evaluation is testing the trained model on S9, S11 and not on S5, S6 etc.

— Reply to this email directly, view it on GitHub https://github.com/mgholamikn/AdaptPose/issues/4#issuecomment-2322787082, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANJS3Z4JVXRDALIDU7FZSF3ZUFLTPAVCNFSM6AAAAABNNN5ZLOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRSG44DOMBYGI . You are receiving this because you were mentioned.Message ID: @.***>