# machine-learning
- u
user
03/20/2021, 10:46 AMIt's a tts software - m
Monero
03/22/2021, 10:33 AMyeah I tried and failed just going to stick with this tts since the model I was training was Eminem and the one you guys releases sound so much better than mine 😄 - z
zwf
03/22/2021, 3:14 PMOh I didn't realize you were working on an eminem voice! - u
user
03/22/2021, 3:14 PMi released a voice dataset - u
user
03/23/2021, 9:37 PMdirectML is something like enviroment - z
zwf
03/23/2021, 9:39 PMreally just the
line and thetf.enable_eager_execution ...
linetf.add() - z
zwf
03/23/2021, 9:39 PMtry running just those in a python console. - u
user
03/23/2021, 9:40 PMnot on anaconda? - u
user
03/23/2021, 9:40 PM? - z
zwf
03/23/2021, 9:42 PMyeah, anaconda - z
zwf
03/23/2021, 9:42 PMsame thing - u
user
03/23/2021, 9:44 PMdo u need to see the sv2tts startup console? - u
user
03/23/2021, 9:44 PMhow i start the sv2tts - u
user
03/23/2021, 9:44 PM? - z
zwf
03/23/2021, 9:45 PMdon't worry about starting sv2tts for now - let's just see if the DirectML setup is working as expected - z
zwf
03/23/2021, 9:46 PMYou should see something like - z
zwf
03/23/2021, 9:46 PMCopy code2020-06-15 11:27:18.235973: I tensorflow/core/common_runtime/dml/dml_device_factory.cc:45] DirectML device enumeration: found 1 compatible adapters. 2020-06-15 11:27:18.240065: I tensorflow/core/common_runtime/dml/dml_device_factory.cc:32] DirectML: creating device on adapter 0 (AMD Radeon VII) 2020-06-15 11:27:18.323949: I tensorflow/stream_executor/platform/default/dso_loader.cc:60] Successfully opened dynamic library DirectMLba106a7c621ea741d2159d8708ee581c11918380.dll 2020-06-15 11:27:18.337830: I tensorflow/core/common_runtime/eager/execute.cc:571] Executing op Add in device /job:localhost/replica:0/task:0/device:DML:0 tf.Tensor([4. 6.], shape=(2,), dtype=float32) - u
user
03/23/2021, 9:46 PMi installed directML and the python script still using my ypu - u
user
03/23/2021, 9:46 PMCPU* - u
user
03/23/2021, 9:46 PMwhat is this? - z
zwf
03/23/2021, 9:47 PMThese are log lines generated from the
call, which tell us that that operation was executed on the DML device. I got it from: https://docs.microsoft.com/en-us/windows/win32/direct3d12/gpu-tensorflow-windowstf.add() - z
- z
zwf
03/25/2021, 1:16 AM^ useful starting point - d
DeletedUser
03/25/2021, 6:19 AMIs there a GitHub for pre trained models that you have? Also is Uberduck using tacotron2? - d
DeletedUser
03/25/2021, 7:05 PM@User - z
zwf
03/25/2021, 7:08 PMIt's using tacotron2 for now, but that might change. I haven't released any pretrained models. - d
DeletedUser
03/25/2021, 7:15 PMOkay thanks! Do you have an estimated time it took you to train steve harvey? Did you only use the steve harvey voice samples .zip plus transcript for training? - z
zwf
03/25/2021, 7:16 PMyep! all my models are fine-tuned on a dataset from half an hour to 5 hours after initial training on LJSpeech. - z
zwf
03/25/2021, 7:16 PMI think Steve is 2 hours or so - d
DeletedUser
03/25/2021, 7:18 PMOkay thanks! I am newer to machine learning as a computer science major. Thanks for the help!