Cant synthesis model with latest master branch

[HenningCode]

Hello Guys,

I got some more issues with the latest master branch. I already got the issue resolved in #515, but now I cant synthesis the model.

The first things that was weird is that the HLS Project is written twice, as seen in the following log.

Interpreting Model
Topology:
Layer name: input_1_m, layer type: InputLayer, input shapes: [[None, 28, 28, 1]], output shape: [None, 28, 28, 1]
Layer name: flatten_1_m, layer type: Reshape, input shapes: [[None, 28, 28, 1]], output shape: [None, 784]
Layer name: qdense_1_m, layer type: QDense, input shapes: [[None, 784]], output shape: [None, 16]
Layer name: qact_1_m, layer type: Activation, input shapes: [[None, 16]], output shape: [None, 16]
Layer name: qdense_2_m, layer type: QDense, input shapes: [[None, 16]], output shape: [None, 128]
Layer name: qact_2_m, layer type: Activation, input shapes: [[None, 128]], output shape: [None, 128]
Layer name: qdense_3_m, layer type: QDense, input shapes: [[None, 128]], output shape: [None, 10]
Layer name: softmax, layer type: Softmax, input shapes: [[None, 10]], output shape: [None, 10]
Creating HLS model
WARNING: Invalid ReuseFactor=2000000 in layer "qdense_1_m". Using ReuseFactor=12544 instead. Valid ReuseFactor(s): 1,2,4,7,8,14,16,28,49,56,98,112,196,392,784,1568,3136,6272,12544.
WARNING: Invalid ReuseFactor=2000000 in layer "qdense_2_m". Using ReuseFactor=2048 instead. Valid ReuseFactor(s): 1,2,4,8,16,32,64,128,256,512,1024,2048.
WARNING: Invalid ReuseFactor=2000000 in layer "qdense_3_m". Using ReuseFactor=1280 instead. Valid ReuseFactor(s): 1,2,4,8,16,32,64,128,256,640,1280.
WARNING: Invalid ReuseFactor=2000000 in layer "qdense_1_m". Using ReuseFactor=12544 instead. Valid ReuseFactor(s): 1,2,4,7,8,14,16,28,49,56,98,112,196,392,784,1568,3136,6272,12544.
WARNING: Invalid ReuseFactor=2000000 in layer "qdense_2_m". Using ReuseFactor=2048 instead. Valid ReuseFactor(s): 1,2,4,8,16,32,64,128,256,512,1024,2048.
WARNING: Invalid ReuseFactor=2000000 in layer "qdense_3_m". Using ReuseFactor=1280 instead. Valid ReuseFactor(s): 1,2,4,8,16,32,64,128,256,640,1280.
Writing HLS project
Done
Writing HLS project
Done

And then if I am trying to build the Project this is the Error I am getting.

ERROR: [HLS 200-70] Part 'None' is not installed.
command 'ap_source' returned error code
    while executing
"source build_prj.tcl"
    ("uplevel" body line 1)
    invoked from within
"uplevel \#0 [list source $arg] "

I am targeting a zcu104 and I know it isnt supported but I used the added board files from @thesps and they worked for me in the past. I also tried to target a pynq, but that results into the same error message.

Any help would be appreciated.

[HenningCode]

This is my Config:

HLSConfig
  LayerName
    input_1_m
      Precision
        result:      ap_fixed<16,6>
    qact_1_m
      Precision
        result:      ap_ufixed<8,4>
      ReuseFactor:   2000000
      Trace:         True
    qact_2_m
      Precision
        result:      ap_ufixed<8,4>
      ReuseFactor:   2000000
      Trace:         True
    qdense_1_m
      Precision
        bias:        ap_fixed<8,3>
        weight:      ap_fixed<8,3>
      ReuseFactor:   2000000
      Trace:         True
    qdense_2_m
      Precision
        bias:        ap_fixed<8,3>
        weight:      ap_fixed<8,3>
      ReuseFactor:   2000000
      Trace:         True
    qdense_3_m
      Precision
        bias:        ap_fixed<8,3>
        weight:      ap_fixed<8,3>
      ReuseFactor:   2000000
      Trace:         True
    softmax
      Precision:     ap_fixed<16,3>
      ReuseFactor:   2000000
      Trace:         True
      exp_table_t:   ap_fixed<18,8,AP_RND,AP_SAT>
      inv_table_t:   ap_fixed<18,8,AP_RND,AP_SAT>
      table_size:    1024
  Model
    Precision:       ap_fixed<16,6>
    ReuseFactor:     2000000
    Strategy:        Resource
Backend:             VivadoAccelerator
ClockPeriod:         10
IOType:              io_stream
OutputDir:           final_training/HLS/fc_small
ProjectName:         fc_small
XilinxPart:          xczu7ev-ffvc1156-2-e
Board:               zcu104
Interface:           axi_stream
InputData:           input_features.dat
OutputPredictions:   output_predictions.dat

Do I need to rename anything maybe ?

Edit:

I saw that the XilinxPart now only called Part.

So I think the Config needs to be created in a different way now. Is there an example how to create the AcceleratorConfig?

[thesps]

I haven't tried out your code, but from your message, I think renaming XilinxPart to Part should fix it. The master branch is undergoing many changes, including to the API, so just be aware that the situation might change again before the next release.

[HenningCode]

Thank you. The only thing unclear atm is how do I set the board aswell? I saw there is a new AcceleratorConfig?

And the Problem with double writing the Writing HLS project still exists?